I remember a network manager telling me the drives on one of our servers were 'hot swappable'. I explained that this was not the case - standing next to me in the server room he said 'look if I pull this disk out the network will stay up'. I spent the next day rebuilding the database he corrupted and the business was without IT systems for the day.
When I was working on mainframe only computer operations staff were allowed in the computer room. With only a few exceptions. Though we had a software raid system so pulling a disk pack wouldn't have bothered it all that much.
Did you know you can break into any mainframe by typing super fast into a green command prompt while techno music plays? It's a law of nature. Abraham Lincoln wrote an essay on it.
Or you can type on two keyboard and multi-task, just like the three main operating systems which are run on an IBM mainframe. And let's bring back CRBE; does anyone remember this?
When she first joined the Navy, there was a war on and they needed ALL HANDS ON DECK, including the manicured ones! Remember, this was the "Rosie the Riveter" era, when the wives and future wives of military men were working overtime building the weapons their men would need to fight the war. But Grace Hopper was a genius in both the mathematics and the human relations of getting high tech projects done. Although she was technically only a "member" of the committee which defined the COBOL language (COmmon Business Oriented Language), her technical expertise was such that most of her suggestions were accepted unanimously by the rest of the committee. She worked her way up to the rank of Rear Admiral (one star rank) before retirement, and became a "nerd celebrity" afterward.
Wonderful and accurate. I've been working on this type of IBM system for 35 years now. This guy "gets them". There are even more redundancies than he mentioned. Automatic cpu recovery, if a CPU fails, a spare takes over and the instruction which died is automatically restarted and the user software doesn't even know it. Same with RAM memory.
What if the instruction was the cause for CPU failure? Wouldn't it bring the whole system down, if a faulty instruction was automatically executed on all backup CPUs?
@@jsimsons On the mainframe, there are no instructions available to directly control hardware at the level you identify. There is a separate x86 server that manages the hardware. The computer operator interacts with the GUI on the x86 box to manage the hardware. In addition, the operating systems run in an LPAR (logical partition) that is a firmware hypervisor.
I have 10 years+ developer experience on IBM mainframes. And I kind of miss the old days. For instance - back then, pressing F3 to save whatever sourcefile I was editing it usually took
A bank I worked for demanded .25 second reaponse time. When we got a new processor I had to throttle back the response times so users would always see rhe .25 response time. But as usual, faster the machine is, the more work they find for it, and eventually my throttle time was zero, tine to get a faster box... COBOL programmer? Which version?
I once visited the IBM datacenter in Böblingen, Germany. It was a truly impressive experience to see hundreds of IBM z13's all together in one big server room. These were probably the most expensive things I've ever seen in one place!
Well, it's called tech *quickie* for a reason. It assumes a level of prior knowledge that's closer to what an average joe would have, hence why certain topics are dumbed down a bit. I mean, the title is "what are mainframes?" and not "How do mainframes work?". Seeing the applications and a broad overview of their functions is way more useful then dwelling on the nitty-gritty.
For the IBM Mainframe affectionados out there. IBM announced the Z15 system, some specs are: up to 40 terabytes (TB) of RAM, up to 1536 I/O channels with over 600 FiConn channels (optical), up to 190 user processors (UP) each with 12 cores at 5.8 GHz, and a slew of specialty processors. The level of complexity can only be described as mind boggling.
"mainframes aren't designed to run games" next up on linus tech tips "1 million dollar main frame smashed in super mario brothers emulator benchmarks by raspberry pi"
Years ago I was at a university that used a burrough b6700 mainframe, and the systems programmer put a very advanced version of the text game "Star trek" on the system. It was a multi player version with secret entry points - programs that did other things, but would transfer to the game when a secret alpha code was typed into a number field. The school administrators didn't like the student using the multi million dollar mainframe for playing games, and had the systems programmer write a monitor program to catch students or staff playing the game. The programmer complied, creating a program call SPY., however an additional feature was added to SPY . Known as Counter SPY, when SPY was running the LOGIN prompt would change to L0GIN and word got out to the game player to not play the game if they saw the zero instead of the oh.
@@tripplefives1402 or through 3270 terminals (real or emulated on PCs). Not graphics oriented but a lot better than line mode (telnet, etc). Or through web pages which can contain graphic images. E.g. www.velocitysoftware.com/zvws.html
Thank you for posting this. As a mainframe systems programmer with over 30 years experience it is nice to see someone with some understanding of what mainframes do. One thing that should be mentioned is that IBM's large systems operating system Z/OS is setup so that it can be 'plexed' together. You can have multiple Z/OS images running across multiple mainframes that are connected in a 'SYSPLEX'. If one system or even one processor is taken down for maintenance the other systems will take up the slack and continue on without the end user even knowing what is going on. Also mainframes do not have internal disk storage. The mainframe is connected to a external storage subsystem like an IBM DS8000 series unit. The disk system is also redundant so that if one of the drives fails it will automatically recover onto another disk drive. The storage subsystem can also be setup so that it takes advantage of asynchronous write. This is where data is written to the the local disk unit as well as an offsite unit at the same time. The disk subsystem can also be encrypted so that all the stored data is totally encrypted at all times.
Thank YOU, DeWitt, for posting your comment. I was a junior systems programmer on an old air-cooled IBM 3083 system in a regional data center, running MVS/XA. It was great while it lasted. (The company ended up consolidating our applications into a bigger system in another city that could run IBM's DB2 database. The limited memory in that 3083 did fine with John Cullinane's IDMS database, but time and software marches on.) When you walked into the machine room with those long strings of cabinets, each with dual 14-inch disk drives run by big three-phase electric motors, you could HEAR data being processed!
I was a system programmer that began on a 704, 1401, 360, 370, 4300 series, 3000 series and finally a s/390 from all sorts of markets. I taught COBOL (D, F & ANSI) Assembler, RPGii and PL/1 (My fav is assembler), taught programmers how to read core dumps & taught mainframe operating sysyems concepts and facilities (DOS & MVS) too at a big college in NYC. I miss my mainframe days. I actually enjoyed it. I got a 370 to do things IBM said wasnt possible. (The code is still running to this day)
@@RaymondHng becuase at the time they did two completely differnet tasks. O/S or MVS was used at very large data centers and used on powerful machines, more adapted to transaction processing. DOS or VSE was used mainly for batch systems even though CICS is used for transaction or online type of stuff. The hardware requirements were a lot more modest. Remember back then you never purchased a mainframe, you rented it. If you really needed a large computer, you would choose MVS type if system anf paid considerably mOre each month. I think differences between MVS and VSE in just JCL alone would be a huge adjustment. They did have VM back then as a means of a crossover from DOS to MVS but it was a real dog as most VM technology is. So people either staued with DOS or had to go through a huge conversion and a retraining process to switch to MVS. It was an expensive proposition. I worked at a shop that did that conversion and it took 3 years to complete. It was a COBOL shop and they had most some if the original COBOL code, so i was brought in to read core dumps, recode it in assembler, test it, than translate the assembler code to COBOL to give back to the programming staff.
As a mainframe system programmer, thanks for clearing up many of the misconceptions and explaining what a mainframe really is so people are better informed :) z/OS FTW :)
As a former IBM SSR, I'm thankful you did a video on this @Linus, I wish more people took interest in enterprise systems like Z. Short and Quick summary, but informative for the length. Thanks!
I started programming on COBOL on System/36 but have moved away from it. Training on mainframe is almost non-existent. Unless your organisation pays IBM to conduct such training.
Cool video. I started as a Cobol programmer on IBM Mainframes in 1979 and worked my way up to Technical Specialist at IBM. Cobol, CICS, DB2, IMS. Left the business in 2004 to do something completely different after working for a major UK retailers on a system that ran over several mainframes and spent all day working out what people will buy in the next 21 days.
I started my IT career as a mainframe operator working in mainframe IBM OS/VM environment, then moved on that latest and greatest IBM MVS/VM. I just about did it all, and you learn on the job from the ground floor up. None of the IT certification stuff, and teamwork counts for everything. Mainframes had very little demand for memory resources and it can far work in multiprocessing much better than the average pc today. Remember the days of the PC that gave the "blue screen of death"? That was unheard of back in the day. No end-users to deal with end-user issues on a regular basis. There was not even a call center or help desk to call or be on wait on the phone. Departmental supervisors were actively involved. No driver installs, and definitely no constant software and operating system updates every month. The pc computer has created such an environmental hazard for the last 40 years. The trash it has produced from floppy and CD/DVD disks, dead disk drives, outdated circut boards, and so forth. Salaries back then were so much better to what the money is equivalent to today. Back then you could get hired and negotiate your salary without much pushback from HR. Outsourcing and offshoring did not happen, but the change of technology and political policies changed everything. Back then, those were better days in the 80s and early 90s.
I know i'll be getting the "old coot" award for this- Back in the 70's, our H.S. used a UNIVAC. You know, vacuum tubes and miles of wires. No monitor or screen, just terminals. (No, not the airport type.) According to our computer lab instructor, that was a mainframe. As was the next year's acquisition of an IBM 360, which most people consider the first mainframe. Computer lab was attached to our vocational electricity/"modern" electronics course, so we, the students of these classes, maintained and repaired the mainframe as part of the course. (COBAL and BASIC, wow.) Anyone else play golf via a terminal? (Then we obtained a P.E.T. - ah, playing the original, "E" and "K" in monochrome, Star Trek.
In the 70s? Thought tubes went out with the 60s. The fitst computer I learned to program was an LGP30 in 1965. It used some tubes but was old then. We programex it in machine code. It hade no high level languages or even an assembler. It only had 64k of 30 bit word memory on a rotating drum.
heh, my old k-12 school had a s/360 low-end mainframe back in the 70's from what I heard. Two Japanese admins at the school programmed it, and I had the honor of learning computer science under one of them (the other taught Assembly in high school as late as 2012), sadly they've retired (which is a shame, I can vouch that Mrs. Miyamoto my CompSci teacher, was an exemplary teacher and mentor). My old school has good CompSci teachers now too, but I don't know if they can carry the torch of their forebears. We also lost our coolest teachers by far, a young hipster programming couple that made a killing in NYC on being really damn good at Ruby on Rails, they introduced me to my first functional programming ever in the form of the then young Clojure, they've moved on to other work.
@@steamerandy i not going to debate what is what. I was only relaying what my H.S. had in those years. Just because you may think tubes went out in the 60s, doesn't mean there weren't mainframes (or even tv sets) that still were running on vacuum tubes.
I've been in mainframes in one capacity or the other for 26 years now, and I've heard that mainframes would be gone by 2010, yet they are are still here and even stronger than they were 26 years ago. The best analogy is that mainframes act as the "freight trains" of the IT world. Processing millions if not billions of transactions 24/7 with zero downtime, whereas server based systems are like the "tractor-trailers" of the IT world. We also use server farms, for internet and network based applications, using Oracle and SQL Server as the main databases. We also use IDMS in some applications on the midrange systems. Our mainframes use IMS and DB2 as the database systems. I'm an IMS database admin myself and have been for the last ten years, before that I was a mainframe applications programmer, working in COBOL and PL/I (a C based language with some COBOL like syntax), when I worked int the manufacturing industry and all the applications are written in COBOL II, Assembler, or in some cases DYL/280 (for quick adhoc reporting). For the most part, companies don't own mainframes, they contract with IBM who, runs the mainframe systems for them at IBM data centers. IBM then provides the operating system, maintenance, data storage, tools and utilities, and systems support per the contract they have with the organization. I may not get to work with all the "pretty colors" of GUI systems, but my job is pretty secure since IMS DBA's are not that common, and I'm one of the youngest members on my team.
They said Cobol was dead already in the late 80:s. Latest fashion then were Unix and C, C++. Where I live in Europe all manufacturing industries moved to other platforms since they are not as transactions intensive as for example banks. DL1, GU, GN , HDAM, HIDAM................ We got a "U0853" abend in production. What to do? Call the database administrators!!! Where I work the lack of skilled IMS DBA's are so severe they recently decided convert all the databases to DB2. Will take some time though.
Will we have a Linus running on a Xeon or a Zen cpu when Linus dies ? How many gigabit cards will he need to do his own number crunching ? Will the improved Linus be able to not drop things ? These are the real things which need to be answered on Techquickie.
--> Fully watch 45 min. video of some nerd dude talking about his mainframe. --> Then watch video about what a mainframe actually is. --> Realize I don't even care about any of this stuff. --> Time well spent. Please kill me.
As a mainframe programmer since 1969 and who wrote millions of line of assembly code, i can tell you that once you code for a mainframe, yoy will never go back to the PC world again. I feel bad for programmer who never had the pleasure or programming or even touching a mainframe. Even though i am a recently retired CIO, and 13 programming languages under my belt, i miss my days on "big iron"
Leon Barsmith It was for sure a very outdated model. Mainframe storage is done on a different device (not the actual mainframe) and only that machine costs more than a million dollars for one of the newest models available at the moment
He got an old one used for cheap from a university. If I remember right it was a z9, which is over ten years old. It wasn't even one of the enterprise models. It didn't have storage and he had to borrow some from a company that sells mainframe support hardware in order to get it running. And he could only afford to run linux on it because the IBM OS licenses are also pretty expensive.
Wow I'm grateful for this video. I had already analyzed the z/architecture and the multitude of unheard of IBM languages (including a couple obscure assembly languages) and why it was designed since the 1960's to virtualize and emulate ancient platforms (that eventually evolved from refrigerator sized components to an on die component we take for granted, e.g. ALU). I just knew they existed recently and had no idea the Y2K was primarily their problem (because they run ancient FORTRAN and COBOL mainly)... Not to say we don't need them but I thought UNIX had replaced them a long time ago, but it's nice to know we actually have entirely different side to computing that's not supercomputers, Intercontinental telecommunications or Corporate DataCenters. I actually wonder if these machines actually don't use the internet (at least primarily) and still have X.25/Frame Relay interconnections.
A Mainframe is like a box of servers, running on one system. It is designed for extreme high availability - so we never power cycle it. It has a trick where we divide memory into two sections (may in some cases be more) and then we defragment the memory, and then shove all the memory in use to one bank, and depower and switch back on the other. The only way to recover lost memory (memory leaks) is to power cycle the memory, but if you have two banks of memory you can power cycle one at a time and never shut the machine down. If you can power cycle both memory banks on a regular basis (like say once a day at 2am) then you never have to switch the machine off. And if you're running / hosting the complete American Express credit network, then you don't ever want that ap to be down for 5 minutes here & there. That's what a mainframe is. It generally has lots of drives and lots of I/O and network connections, redundant power supplies and drives and almost everything, CPUs can be hot-swapped, memory modules can be hot-swapped, drives can be hot-swapped, whole cabinets and back-planes can be hot swapped ... and they run a very specialised operating system, although IBM has taken a great interest in Linux, and many mainframes today run either a Linux variant inside an old style big iron O/S, or the other way around. The principle feature of a mainframe, is it never gets switched off. Once it has been brought up, it stays up until it is retired.
great one! another interesting fact about them is that their availability can be improved even further with a Geographically Dispersed Parallel Sysplex, a cluster among servers on different cities that will redirect traffic to another server in case of a site failure.
Theoretically yes... They run linux! I have yet to find source for Crysis to compile it on s390x on my machine... But Minecraft crys when trying to run that so...
I used to work at a company that WAS NOT IBM, that built mainframes, some of them were the large rack type that were 6 or 8 feet tall, and FILLED with boards and processors, other mainframes were small rack mount types that were used in aircraft, and ships that like you said were for redundant functions!! We even built the related boards and hardware that went into those mainframes as well!!! And part of my job was "box building" which was actually putting the wire harnesses and the metal cabinets themselves together!! Sorry for those folks who hate people who edit their posts....but I had to add this here!!! Linus you forget mainframes are NOT only used at stores or other traditional businesses! Mainframes IN FACT are still a major part of places like NASA sure....BUT they also run Google and Amazon as well as Ebay, so there are a lot of places that use mainframe computers, and at a different company I worked at we built printers for companies like Target retail stores, and we had a room in the build known as the "IT Jungle" that was about 20 foot by 20 foot room with probably 30 mainframes STUFFED in it, that were all daisy chained together for our "in house" computer network AND print function back up!! In fact ONE of the mainframes was set up as our own personally "in house" server, and with a few clicks of a mouse on ANY computer, they could update ALL the software to the mainframe that ran all the other networked printer test computers!!
Editing your post is a good thing. Nothing wrong with editing posts to make them accurate/clearer/more complete. You might rethink the repetitive use of of multiple exclamation marks at the end of each sentence. They make it less likely that your comment will be taken seriously.
The second picture you used when referring to modern airliners and redundant systems was the cockpit of a DA-40 or Diamond Star. That's a small, single engine training aircraft that doesn't really have any redundant systems.
I once worked for a place that had a Stratus computer. It had three banks: Memory, CPU, and I/O. Each card was paired. (There were like a dozen of more cards in each bank.) If any of the cards died, say a CPU card, its partner was already doing the same processing.The computer would automatically call Stratus Corp, who would overnight you a new card. It would be on your desk in the morning. You take it to the Stratus. Open the big door. Find the card with the red light. Pull it out. Slide the new one in. In literally just a few seconds the green light comes on and you're done! (Well, close the door, too.) Zero down time. The users never know. I find this totally amazing. Can you swap your CPU out with the machine still up and running? Very cool. But also very expensive. I guess this was like mini-mainframe, but it's primary mission was fault tolerance. Everything was in pairs except the backplane and the tape drive. Disclaimer: I do not work for Stratus. I do not get any compensation whatsoever for posts like this -- or from anyone, AAMAF.
One of the very few rewarding experiences of working at Walmart was getting to see the store's mainframe. I wasn't sure what it was at first but it's still the only of I've ever seen.
You know what really would have made this video more valuable? Contrasting mainframes from data centers, which share a lot of the same properties on more scalable (mostly) commodity hardware. :P
I programmed on a mini-main-frame, it had 256 terminals. One day in the computer room the disk drive (size of a washing machine) had skidded across the room because the read/write heads were processing 10000s of transactions and moving so fast, we bolted the drive to the floor. Luckily the cable held and there was no problems.
I LOVED your video about mainframes. Well summarized. And actually, I saw one /360 working! I am a mainframe specialist, and although I am finding difficult to find another project to work on, I still love mainframes!
Used to do work using a mainframe, back in the 80s, at a university. Then it was Macs and PCs. I had no idea that they were still in use until I saw this video.
+Authentically Authentic It's really not that bold. All they have to do is throw WireShark, which is free, onto their network and have it log all the data that comes in.
whether they can or not is a moot point, wireshark/similar tools is taught in basic networking classes not to mention much more automatable methods should be easy for the engineers tunnel bear has. The question is whether or not they do it and like linus says in I think every full length sponsor spot for them, they have a top notch privacy policy that is VERY descriptive about the little data they do collect and what they do with said data, hell they don't even ask for your last name anymore because privacy -edit and besides this way we can spite our ISPs for trying to sell our data by giving them a bunch of useless junk
+Jon Rueckert As much as in all honesty I don't actually think they collect private data, (A) collecting your last name isn't really relevant to Internet packet sniffing, and (B) what's preventing them from simply ignoring their own publicly stated policy?
i could be remembering this wrong but does a company not have to abide by its privacy policy by law in canada (which tunnel bear is based in) not to mention the really bad press
Biggest difference is that mainframes runs one of few big instances (TB of size in RAM). For example data from credit cards around the world. This is why it must be centralized in one machine and with such a overengineered throughput. You don't want failure of your money transaction because of incorrectly updated database. On the other hand, supercomputers usually splits problem between nodes and then just collect results. Since that all node runs independently. So computers don't really knows what how is computers around them doing. For example. If you upload youtube video in Paris it will take some time until they move it to datacenter in west US so local viewers can access it. In case of youtube video it is not & problem, but it might be problem if your customer make one credit transaction in Paris and right after second in LA, before US datacenter is updated.
A small point of detail. I work on systems like this for a living. The Term "mainframe" is technically outmoded. New mainframe are actually designated System Z. The newest model being the Z13. So the proper term here is IBM system Z. HP equivalent (I use this term very loosely) would be the now defunct "Superdome 2"
As someone who worked with a variety of IBM and Wang Mainframes in the 1970's and 1980's I feel well qualified to answer this question. In 1974 the IBM computers were as big as a small truck with extra equipment attached that was often the size of a dozen small cars, They filled a medium warehouse sized space and required air conditioning and three phase power to operate. The hard drives had a capacity of 30 Megabyte and each unit was the size of a washing machine and the disks which were removable weighed around 40 pounds each and were the size of a small car tyre. As for the WANG equipment in 1980 it was about a quarter of the space and had lost many of the additional equipment pieces but still used hard drives of a capacity of 50Mb and about half the thickness of a small car tyre. In the IBM the total memory capacity of the computer was about 64Kb while the WANG unit was about 256KB. By comparison your smartphone is approximately 1000 times faster and 100,000 times smaller.
You mean the Wang VS midrange line? I had to fight tooth and nail to keep them out of our department. The parent company wanted to force all the departments to convert over to Wang and we did not want to spend the money to go through a conversion nightmare.
I was skeptical of this video, but this is actually a pretty good explanation. I write firmware for these things and I usually have a pretty hard time explaining it to lay people.
Haha your cute. Of-course it can! The newer ones anyway. The mainframe runs Linux and has Java. I have put a Minecraft server on my 2004 era mainframe and it ran... but was too slow...
In the late 90s the Cleveland airport still had vintage plain text computer monitors showing flights to travelers and large dot matrix printers at gates. Y2K and 9/11 really sped up upgrades, but in truth that old stuff was probably working just fine.
I worked on (programmed and lead software design teams) for “Big Iron” mainframes (360/370, etc) for almost 40 years. When mini computers (which didn’t last) started to come on the scene, talk was that they would “replace MFs.” I had a DEC mini-computer person call me up to understand how we solved a very complicated statistical problem on a MF. I spent about 5 minutes describing the process we used (which involved creating millions of extracts to be sorted to inflate our small amount of sample date to project the results at 4 organizational levels of the universe of data answers). When I finished I asked him, “Can you do that on a mini computer?” All I heard was silence and then he thanked me before he hung up. I saw a picture of a T-Rex chewing up some mashed metal and the caption said, “Mainframes are back and they are Mad!” I used to say that PC’s are fast but Mainframes SCREAM! And look how far we’ve come with “Big Iron” MFs since then. Mainframes SCREAM!
Actually more and more financial institutions either abandon zOs completely to more container based processing (microservices, DC/OS, Docker), even moving their DB2 off Mainframe and going the Oracle route or they move to Linux VMs on z/OS and run their J2EE services on that. This process has started in the early 2000s when I was working for IBM, actually a few national banks actually moved from Mainframe to Windows in Azure! (Yeah I know crazy hey!) The biggest problem is that native z/OS developers are hard to find and C/C++/COBOL/REXX are not really suited for the highly complex integrated solutions we have today. Business logic can become so enormously complex and distributed that you need an Enterprise Framework to help you cope with this madness. Most big banks here in NL have a host of J2EE solutions for their enterprise software infront of their mainframe and the mainframe just does the financial transactions offered from the Enterprise software. I know for a fact because I helped build a continuous deployment solution. Also you can look at the terminal at an airport but most airlines have a screenscraping/GUI interface over the SITA system.
" the mainframe just does the financial transactions offered from the Enterprise software."... which is what it's designed to do. If the migration process started in the early 2000s, and 96 of the world's top 100 banks and 23 of the top 25 retailers are still using them, then I'd say they realized that a rats nest of microservices wasn't up to the task of replacing their big iron.
bingo. At least those companies that want to be competitive in today's world. ; ). Actually its surprising how many are migrating to cloud/VM architecture. McDonalds, GE etc. The screen scrape thing makes me want to hurl.
That green screen you see might also be IBM i (AS/400) also still alive and well since 1987. For pure transactional power, the largest IBM Power9 system running IBM i (2.74M CPW on 192 cores) now bests the largest Z15 (1.25M CPW on 190 cores) (MIPS converted to CPW). Power10 will also beat Z16 core/core.
Yeah. I am studying in Japan and I visited Japan's fastest super computer "Kei" in Kobe. The processors are made by Fujitsu and they are interconnected via their proprietary MPI system. Each processors is directly connected with 6 processors in the network in like 3 dimension like the atomic structure of a crystal. I forgot the exact name of that connection but it sounds like 3D blah blah blah.
I am glad you created this video. I used to work on a pair of mainframes at a large global car manufacturer and I assisted there with the mainframes SNA connectivity amongst other things (SNA is kind of like TCP/IP but for mainframes). Mainframes are often kicked in the teeth because PC's obviously outsell them, are ubiquitous and few have true experience with them in order to make a realistic comparison to a PC. As you say, they are architecturally almost completely different and built from the ground up to work with a different kind of load to what is PC is meant for. Because of a mainframe's architecture, it can support hundreds if not thousands of users with a 1/10th of the RAM a PC would need for it to battle to cope with just one neurotic high-end user. But of course, a mainframe ain't built for making pretty pictures like a PC is. These machines live in a different but parallel dimension .
AS/400 is just enterprise software, and runs on any reasonably powerful (preferably highly multi-thread)CPU. We ran it on an older xeon chip where I work, and just replaced it 2 years ago...
The AS/400 was intended for small shops and supported job running protocols that were not compatible with the "big iron" mainframes in the Syatem 360/370/390 series.
"The IBM Mainframe Manual", if printed on actual paper, small font, would not fit in the bank vault of the home office of Citibank if you cleared out all the money, lock boxes, gold bricks, and fire protection. These machines which some of you write off as uninteresting and irrelevant hold 70% of the world's data (including much of the "Big Data"), run in 97% of the "Fortune 1000" companies, and generally make your comfortable life POSSIBLE. You don't spend a dime with a credit card, book a flight, hotel, or car rental, communicate with your cell phone and/or any other small device (think laptops, tablets, etc) without what you're typing passing through, and likely being at least processed at some point by one (or more) mainframes. Whether you're interested or not, most of your life would not be even remotely as comfortable, functional, organized, easy, or safe without these machines (which of course are used to design, manufacture, distribute, and finance the cars you drive (whether purchased, leased, rented, Uber'ed or somehow otherwise shared). And, as noted, one mainframe might host (i.e. RUN) 8000 images of "real" (as defined by non-mainframers) "Intel (or other) server" machines (that all of you think of as "big machines") under a Virtual Machine (Hypervisor) running on PART of a single mainframe, along with other zOS or other operating system workloads: all this in ONE MAINFRAME. So, dismiss them as you like. Or appreciate them as a large part of every person's life in any civilized part of the planet. - "Proud Mainframer, since 1968..."
IBM Mainframes supported Graphics Terminals using software called GDDM. The API was similiar to OS/2 Presentation Manager GPI. Actually GDDM came first, so GPI was probably based on it. Mainframe graphics were not very widely used.
Before that you had Programmed Symbols, where you could make up up to 6 different fonts (9 pixels high by 16 wide) of 192 characters each. They could be single-plane (one colour) or triple plane (where you have red/green/blue planes).
Linus my professor of economics says that desktop pc's are going to die in the future because our phones are getting more powerful, what do you think about that? Personally I think he's wrong but I can't give him an specifics reason. For example if the reason is video editing, it isn't a big enough market to keep the industry going. Thanks
PCs have their purpose just like phones do, and they are both different. Phones will always have the use of being the on-the-go, less powerful pieces of hardware where you can accomplish many everyday needs. PCs will always have the use of being the more potent pieces of hardware where you accomplish what simply cannot be done on a phone. The thing is, eventually we will reach a point where we cannot make our hardware any smaller to make space for power, and from that point on, we will only have to make things bigger, and as for phones, eventually they will reach their peak height and width (that which is the maximum height and width a phone can retain while still being comfortable and small enough to carry in your case, pants, purse, etc. Desktop PCs can continue to maintain their size as larger pieces of hardware with tight but large and potent parts inside. Therefore, PCs will ALWAYS be more powerful than phones, allowing them to not die out.
Felipe Prieto for consumers, yes, not laptops, but low power towers yes, maybe some will still exist for ofices and shait, but high power desktops? they will never go away, its just a fact, because of how fast they get upgrades, top of the line phones, consoles and laptops will never be faster than a top of the line desktop, simply because it takes way more time to make a complete system than just a single component to plug on a desktop, so yeah, desktops will never go away completely.
Felipe Prieto Desktop computers serve plenty of purpose, and will continue to do so. You cannot program effectively on a phone or tablet. A physical keyboard, phone realestate and higher thermal dissipation mean desktops are the more suited form for power-hungry tasks. Even just multi-tasking can be a pain on mobile devices.
Authentically Authentic right under the video on android i can, upvote or downvote the video, share, "save", or and to playlist. mostly referencing the save button.
I started working with mainframes a few years ago and am still stumped on explaining what I do for work. I think this will actually help explaining my job a bit better. Thoug it still ends with "I get paid for keeping it alive". lol
Early mainframes, and I worked on some in the 70's couldn't just be turned on. You had to power it up, hit a set of switches and press execute several times to give it the instructions to read a tape strip or cassette. Then that info would let it read a 7 or 9 track tape that let it read the OS tape. Older systems would read cards.
@Richard Conte in the real old days you turned power on, then presses switches that entered a short tape or card read program . The system then read about 20 cards or a short tape that gave it a very short set of instructions. Then load a larger tape or stack of cards that held the OS then you typed xqt and tape/card identifier it would the n load the operating system. You then loaded the COBOL, FORTRAN systems, then you ran your programs such as payroll, parts ordering, etc. It was time consuming, you may only have 1,024 bytes of memory and storage was maybe 640K bytes on a drum.
I remember a network manager telling me the drives on one of our servers were 'hot swappable'.
I explained that this was not the case - standing next to me in the server room he said 'look if I pull this disk out the network will stay up'.
I spent the next day rebuilding the database he corrupted and the business was without IT systems for the day.
Haha. I've known a lot of managers like that!
*Facepalm*
That's why he's a network manager.
When I was working on mainframe only computer operations staff were allowed in the computer room. With only a few exceptions. Though we had a software raid system so pulling a disk pack wouldn't have bothered it all that much.
This is why he's a manager. It's called he's giving you work to do. :)
Did you know you can break into any mainframe by typing super fast into a green command prompt while techno music plays? It's a law of nature. Abraham Lincoln wrote an essay on it.
This is correct.
:-)
And as NCIS has shown, two people typing on halves of the same keyboard can hack faster! Four hands is better than two, right?
Trenchcoat and shades give 50% speed increase and each smash of enter button reduces the remaining time by 30 seconds!
Or you can type on two keyboard and multi-task, just like the three main operating systems which are run on an IBM mainframe. And let's bring back CRBE; does anyone remember this?
As a mainframe COBOL programmer I greatly appreciate this video! Yes, I was born after 1980.
amjohnson86 i learned cobol in college. and im 27, graduated 4 years ago.
We have some COBOL programmers in my company, it's like taking a trip into the past. No way I would actually write code in that, though.
Dr. Grace Murray Hopper, yes a woman was responsible for the creation of COBOL. How did the evil patriarchy let that happen?
When she first joined the Navy, there was a war on and they needed ALL HANDS ON DECK, including the manicured ones! Remember, this was the "Rosie the Riveter" era, when the wives and future wives of military men were working overtime building the weapons their men would need to fight the war.
But Grace Hopper was a genius in both the mathematics and the human relations of getting high tech projects done. Although she was technically only a "member" of the committee which defined the COBOL language (COmmon Business Oriented Language), her technical expertise was such that most of her suggestions were accepted unanimously by the rest of the committee. She worked her way up to the rank of Rear Admiral (one star rank) before retirement, and became a "nerd celebrity" afterward.
Woody Can I ask what college you went to that is teaching COBOL? I have been trying to find some college level courses in COBOL for my work.
Wonderful and accurate. I've been working on this type of IBM system for 35 years now. This guy "gets them". There are even more redundancies than he mentioned. Automatic cpu recovery, if a CPU fails, a spare takes over and the instruction which died is automatically restarted and the user software doesn't even know it. Same with RAM memory.
What if the instruction was the cause for CPU failure? Wouldn't it bring the whole system down, if a faulty instruction was automatically executed on all backup CPUs?
Its like RAID or having a hot spare on RAM and CPU! Now thats hot!
@@jsimsons WTF! What instruction could cause failure? Do you think that in the instruction set there is a "die" command?
"RAM memory" aaarggghhhh
@@jsimsons On the mainframe, there are no instructions available to directly control hardware at the level you identify. There is a separate x86 server that manages the hardware. The computer operator interacts with the GUI on the x86 box to manage the hardware. In addition, the operating systems run in an LPAR (logical partition) that is a firmware hypervisor.
3:40 the lady's hand aligns with eye
You get a thumb.
GOT EEEMMM
And yours too!
Illuminati confirmed
I thought she was a Klingon.
I have 10 years+ developer experience on IBM mainframes. And I kind of miss the old days. For instance - back then, pressing F3 to save whatever sourcefile I was editing it usually took
A bank I worked for demanded .25 second reaponse time. When we got a new processor I had to throttle back the response times so users would always see rhe .25 response time. But as usual, faster the machine is, the more work they find for it, and eventually my throttle time was zero, tine to get a faster box...
COBOL programmer? Which version?
I agree I miss those days too. Most of all NO END-USERS.
"Big Iron" just reminds me of the Fallout: New Vegas soundtrack.
Ben Toth: Thought I was the only one.
Ben Toth Ah.. I see you are a man of culture.
clearly the only correct answer
biiiiiiiiiig irrrrrrrrrrrrron biiiiiiiiiiiiiiig irrrrrrrrrrrrrrron
big iron on his hiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiip.
I once visited the IBM datacenter in Böblingen, Germany. It was a truly impressive experience to see hundreds of IBM z13's all together in one big server room. These were probably the most expensive things I've ever seen in one place!
And here I thought I knew what mainframes were.
Well, it's called tech *quickie* for a reason. It assumes a level of prior knowledge that's closer to what an average joe would have, hence why certain topics are dumbed down a bit. I mean, the title is "what are mainframes?" and not "How do mainframes work?". Seeing the applications and a broad overview of their functions is way more useful then dwelling on the nitty-gritty.
For you, maybe. For me, I found it quite good mainly because I (like many others) didn't think of mainframes as pieces of modern technology.
So tell me now, what are mainframes? ;) :D
I thought they were cold war computers used to launch nuclear missiles
Me too
For the IBM Mainframe affectionados out there. IBM announced the Z15 system, some specs are: up to 40 terabytes (TB) of RAM, up to 1536 I/O channels with over 600 FiConn channels (optical), up to 190 user processors (UP) each with 12 cores at 5.8 GHz, and a slew of specialty processors. The level of complexity can only be described as mind boggling.
Can it run doom
@@Askial_Osial You can run a minecraft server on a mainframe
You can have hundreds of VMs with all type of OS running on a mainframe. So, yes, DOS and Doom would work. 😅
"mainframes aren't designed to run games" next up on linus tech tips "1 million dollar main frame smashed in super mario brothers emulator benchmarks by raspberry pi"
Yeah but can run text games like adventure, etc. on VM system. UNIX and VMS are designed to run graphical games.
en.wikipedia.org/wiki/Gameframe
Years ago I was at a university that used a burrough b6700 mainframe, and the systems programmer put a very advanced version of the text game "Star trek" on the system. It was a multi player version with secret entry points - programs that did other things, but would transfer to the game when a secret alpha code was typed into a number field.
The school administrators didn't like the student using the multi million dollar mainframe for playing games, and had the systems programmer write a monitor program to catch students or staff playing the game. The programmer complied, creating a program call SPY., however an additional feature was added to SPY . Known as Counter SPY, when SPY was running the LOGIN prompt would change to L0GIN and word got out to the game player to not play the game if they saw the zero instead of the oh.
@@garyclouse4164 wow
@@tripplefives1402 or through 3270 terminals (real or emulated on PCs). Not graphics oriented but a lot better than line mode (telnet, etc). Or through web pages which can contain graphic images. E.g. www.velocitysoftware.com/zvws.html
Thank you for posting this. As a mainframe systems programmer with over 30 years experience it is nice to see someone with some understanding of what mainframes do. One thing that should be mentioned is that IBM's large systems operating system Z/OS is setup so that it can be 'plexed' together. You can have multiple Z/OS images running across multiple mainframes that are connected in a 'SYSPLEX'. If one system or even one processor is taken down for maintenance the other systems will take up the slack and continue on without the end user even knowing what is going on. Also mainframes do not have internal disk storage. The mainframe is connected to a external storage subsystem like an IBM DS8000 series unit. The disk system is also redundant so that if one of the drives fails it will automatically recover onto another disk drive. The storage subsystem can also be setup so that it takes advantage of asynchronous write. This is where data is written to the the local disk unit as well as an offsite unit at the same time. The disk subsystem can also be encrypted so that all the stored data is totally encrypted at all times.
Thank YOU, DeWitt, for posting your comment. I was a junior systems programmer on an old air-cooled IBM 3083 system in a regional data center, running MVS/XA. It was great while it lasted. (The company ended up consolidating our applications into a bigger system in another city that could run IBM's DB2 database. The limited memory in that 3083 did fine with John Cullinane's IDMS database, but time and software marches on.) When you walked into the machine room with those long strings of cabinets, each with dual 14-inch disk drives run by big three-phase electric motors, you could HEAR data being processed!
I was a system programmer that began on a 704, 1401, 360, 370, 4300 series, 3000 series and finally a s/390 from all sorts of markets. I taught COBOL (D, F & ANSI) Assembler, RPGii and PL/1 (My fav is assembler), taught programmers how to read core dumps & taught mainframe operating sysyems concepts and facilities (DOS & MVS) too at a big college in NYC.
I miss my mainframe days. I actually enjoyed it. I got a 370 to do things IBM said wasnt possible. (The code is still running to this day)
@@rty1955 So why did IBM never merge z/OS and z/VSE into one operating system?
@@RaymondHng becuase at the time they did two completely differnet tasks. O/S or MVS was used at very large data centers and used on powerful machines, more adapted to transaction processing. DOS or VSE was used mainly for batch systems even though CICS is used for transaction or online type of stuff. The hardware requirements were a lot more modest. Remember back then you never purchased a mainframe, you rented it. If you really needed a large computer, you would choose MVS type if system anf paid considerably mOre each month.
I think differences between MVS and VSE in just JCL alone would be a huge adjustment. They did have VM back then as a means of a crossover from DOS to MVS but it was a real dog as most VM technology is. So people either staued with DOS or had to go through a huge conversion and a retraining process to switch to MVS. It was an expensive proposition.
I worked at a shop that did that conversion and it took 3 years to complete. It was a COBOL shop and they had most some if the original COBOL code, so i was brought in to read core dumps, recode it in assembler, test it, than translate the assembler code to COBOL to give back to the programming staff.
As a mainframe system programmer, thanks for clearing up many of the misconceptions and explaining what a mainframe really is so people are better informed :)
z/OS FTW :)
Throws down the z/VM card that's running your z/OS.
Test Plan Charlie FTW.
@@stonent Virtual sysplex for those who only have one CPU, not a CPC.
@@stonent Best answer I've seen in a long time.
(46 year IBM mainframe, mostly VM sysprog)
I agree totally z/os
As a former IBM SSR, I'm thankful you did a video on this @Linus, I wish more people took interest in enterprise systems like Z. Short and Quick summary, but informative for the length. Thanks!
I started programming on COBOL on System/36 but have moved away from it. Training on mainframe is almost non-existent. Unless your organisation pays IBM to conduct such training.
Shame schools no longer teach it. I taught mainframe operarting systems concepts & facilities for years (both DOS/VSE & MVS)
Seriously, one of your best videos to date. I thought everyone forgot about Mainframes. Kudos for the topic.
Cool video. I started as a Cobol programmer on IBM Mainframes in 1979 and worked my way up to Technical Specialist at IBM. Cobol, CICS, DB2, IMS. Left the business in 2004 to do something completely different after working for a major UK retailers on a system that ran over several mainframes and spent all day working out what people will buy in the next 21 days.
Linus when I hear "big iron" I think a heavy caliber six shooter.
Thanks for being mostly correct (latest IBM is the z14 ;-) ). When I hear "Big Iron" I think of a very satisfying 46 year career.
I started my IT career as a mainframe operator working in mainframe IBM OS/VM environment, then moved on that latest and greatest IBM MVS/VM. I just about did it all, and you learn on the job from the ground floor up. None of the IT certification stuff, and teamwork counts for everything. Mainframes had very little demand for memory resources and it can far work in multiprocessing much better than the average pc today. Remember the days of the PC that gave the "blue screen of death"? That was unheard of back in the day. No end-users to deal with end-user issues on a regular basis. There was not even a call center or help desk to call or be on wait on the phone. Departmental supervisors were actively involved. No driver installs, and definitely no constant software and operating system updates every month. The pc computer has created such an environmental hazard for the last 40 years. The trash it has produced from floppy and CD/DVD disks, dead disk drives, outdated circut boards, and so forth. Salaries back then were so much better to what the money is equivalent to today. Back then you could get hired and negotiate your salary without much pushback from HR. Outsourcing and offshoring did not happen, but the change of technology and political policies changed everything. Back then, those were better days in the 80s and early 90s.
z16 is now coming out with the Telum processor, my job as a COBOL programmer has never been more secure
I know i'll be getting the "old coot" award for this- Back in the 70's, our H.S. used a UNIVAC. You know, vacuum tubes and miles of wires. No monitor or screen, just terminals. (No, not the airport type.)
According to our computer lab instructor, that was a mainframe. As was the next year's acquisition of an IBM 360, which most people consider the first mainframe. Computer lab was attached to our vocational electricity/"modern" electronics course, so we, the students of these classes, maintained and repaired the mainframe as part of the course. (COBAL and BASIC, wow.)
Anyone else play golf via a terminal? (Then we obtained a P.E.T. - ah, playing the original, "E" and "K" in monochrome, Star Trek.
In the 70s? Thought tubes went out with the 60s. The fitst computer I learned to program was an LGP30 in 1965. It used some tubes but was old then. We programex it in machine code. It hade no high level languages or even an assembler. It only had 64k of 30 bit word memory on a rotating drum.
Haha yoy forgot the 1401 and the 704. Thte 1401 was rhe first mainframe. I began working on a 407, then 1401 and up the line to a s/390
heh, my old k-12 school had a s/360 low-end mainframe back in the 70's from what I heard. Two Japanese admins at the school programmed it, and I had the honor of learning computer science under one of them (the other taught Assembly in high school as late as 2012), sadly they've retired (which is a shame, I can vouch that Mrs. Miyamoto my CompSci teacher, was an exemplary teacher and mentor). My old school has good CompSci teachers now too, but I don't know if they can carry the torch of their forebears. We also lost our coolest teachers by far, a young hipster programming couple that made a killing in NYC on being really damn good at Ruby on Rails, they introduced me to my first functional programming ever in the form of the then young Clojure, they've moved on to other work.
@@steamerandy i not going to debate what is what. I was only relaying what my H.S. had in those years. Just because you may think tubes went out in the 60s, doesn't mean there weren't mainframes (or even tv sets) that still were running on vacuum tubes.
I've been in mainframes in one capacity or the other for 26 years now, and I've heard that mainframes would be gone by 2010, yet they are are still here and even stronger than they were 26 years ago. The best analogy is that mainframes act as the "freight trains" of the IT world. Processing millions if not billions of transactions 24/7 with zero downtime, whereas server based systems are like the "tractor-trailers" of the IT world. We also use server farms, for internet and network based applications, using Oracle and SQL Server as the main databases. We also use IDMS in some applications on the midrange systems. Our mainframes use IMS and DB2 as the database systems. I'm an IMS database admin myself and have been for the last ten years, before that I was a mainframe applications programmer, working in COBOL and PL/I (a C based language with some COBOL like syntax), when I worked int the manufacturing industry and all the applications are written in COBOL II, Assembler, or in some cases DYL/280 (for quick adhoc reporting). For the most part, companies don't own mainframes, they contract with IBM who, runs the mainframe systems for them at IBM data centers. IBM then provides the operating system, maintenance, data storage, tools and utilities, and systems support per the contract they have with the organization. I may not get to work with all the "pretty colors" of GUI systems, but my job is pretty secure since IMS DBA's are not that common, and I'm one of the youngest members on my team.
They said Cobol was dead already in the late 80:s. Latest fashion then were Unix and C, C++. Where I live in Europe all manufacturing industries moved to other platforms since they are not as transactions intensive as for example banks.
DL1, GU, GN , HDAM, HIDAM................ We got a "U0853" abend in production. What to do? Call the database administrators!!! Where I work the lack of skilled IMS DBA's are so severe they recently decided convert all the databases to DB2. Will take some time though.
Will we have a Linus running on a Xeon or a Zen cpu when Linus dies ? How many gigabit cards will he need to do his own number crunching ? Will the improved Linus be able to not drop things ? These are the real things which need to be answered on Techquickie.
This is you's guys' most important channel, IMO.
--> Fully watch 45 min. video of some nerd dude talking about his mainframe.
--> Then watch video about what a mainframe actually is.
--> Realize I don't even care about any of this stuff.
--> Time well spent.
Please kill me.
You could do a lot worse with your time.
One year later the cycle repeats itself.
That dude who did that lecture has the most supportive father in history, it must be.
dude, sameeeeeeee
Just saw that video too
As a mainframe programmer since 1969 and who wrote millions of line of assembly code, i can tell you that once you code for a mainframe, yoy will never go back to the PC world again.
I feel bad for programmer who never had the pleasure or programming or even touching a mainframe. Even though i am a recently retired CIO, and 13 programming languages under my belt, i miss my days on "big iron"
How about that 18 year old who bought a mainframe for like, $300?
Leon Barsmith It was for sure a very outdated model. Mainframe storage is done on a different device (not the actual mainframe) and only that machine costs more than a million dollars for one of the newest models available at the moment
He got an old one used for cheap from a university. If I remember right it was a z9, which is over ten years old. It wasn't even one of the enterprise models. It didn't have storage and he had to borrow some from a company that sells mainframe support hardware in order to get it running. And he could only afford to run linux on it because the IBM OS licenses are also pretty expensive.
Did that kid ever get a decent haircut?
Leon Barsmith the main frames are basically free when you look at the prices of licensing processes.
He actually works on the mainframe here at IBM now :)
I love it when you answer questions I didn't know I had
Wow I'm grateful for this video. I had already analyzed the z/architecture and the multitude of unheard of IBM languages (including a couple obscure assembly languages) and why it was designed since the 1960's to virtualize and emulate ancient platforms (that eventually evolved from refrigerator sized components to an on die component we take for granted, e.g. ALU). I just knew they existed recently and had no idea the Y2K was primarily their problem (because they run ancient FORTRAN and COBOL mainly)...
Not to say we don't need them but I thought UNIX had replaced them a long time ago, but it's nice to know we actually have entirely different side to computing that's not supercomputers, Intercontinental telecommunications or Corporate DataCenters. I actually wonder if these machines actually don't use the internet (at least primarily) and still have X.25/Frame Relay interconnections.
A Mainframe is like a box of servers, running on one system. It is designed for extreme high availability - so we never power cycle it. It has a trick where we divide memory into two sections (may in some cases be more) and then we defragment the memory, and then shove all the memory in use to one bank, and depower and switch back on the other. The only way to recover lost memory (memory leaks) is to power cycle the memory, but if you have two banks of memory you can power cycle one at a time and never shut the machine down. If you can power cycle both memory banks on a regular basis (like say once a day at 2am) then you never have to switch the machine off. And if you're running / hosting the complete American Express credit network, then you don't ever want that ap to be down for 5 minutes here & there. That's what a mainframe is. It generally has lots of drives and lots of I/O and network connections, redundant power supplies and drives and almost everything, CPUs can be hot-swapped, memory modules can be hot-swapped, drives can be hot-swapped, whole cabinets and back-planes can be hot swapped ... and they run a very specialised operating system, although IBM has taken a great interest in Linux, and many mainframes today run either a Linux variant inside an old style big iron O/S, or the other way around. The principle feature of a mainframe, is it never gets switched off. Once it has been brought up, it stays up until it is retired.
I don't know why, but I have fond memories of working on monochrome terminals connected to Vax/Vms mainframes. The Vax OS was fun and quite powerful.
It was great and DCL made more sense than a lot of scripting languages today
great one! another interesting fact about them is that their availability can be improved even further with a Geographically Dispersed Parallel Sysplex, a cluster among servers on different cities that will redirect traffic to another server in case of a site failure.
Been doing that for decades
but, can they run crysis?
KraineK xDDD karambitch minesweeper any one?
If they dont run, someone will be on a crisis...
Theoretically yes... They run linux! I have yet to find source for Crysis to compile it on s390x on my machine... But Minecraft crys when trying to run that so...
Literally said it cannot play games very well (idc that it's a joke)
have you seen the part on 4:56
Thank you!!
I get so many people asking me what a mainframe is or even if we still use them. I am sending people to watch your vid😊
No, the correct correlation to "Big Iron" is Marty Robbins and the big iron on his hip.
I used to work at a company that WAS NOT IBM, that built mainframes, some of them were the large rack type that were 6 or 8 feet tall, and FILLED with boards and processors, other mainframes were small rack mount types that were used in aircraft, and ships that like you said were for redundant functions!! We even built the related boards and hardware that went into those mainframes as well!!! And part of my job was "box building" which was actually putting the wire harnesses and the metal cabinets themselves together!!
Sorry for those folks who hate people who edit their posts....but I had to add this here!!! Linus you forget mainframes are NOT only used at stores or other traditional businesses! Mainframes IN FACT are still a major part of places like NASA sure....BUT they also run Google and Amazon as well as Ebay, so there are a lot of places that use mainframe computers, and at a different company I worked at we built printers for companies like Target retail stores, and we had a room in the build known as the "IT Jungle" that was about 20 foot by 20 foot room with probably 30 mainframes STUFFED in it, that were all daisy chained together for our "in house" computer network AND print function back up!! In fact ONE of the mainframes was set up as our own personally "in house" server, and with a few clicks of a mouse on ANY computer, they could update ALL the software to the mainframe that ran all the other networked printer test computers!!
Editing your post is a good thing. Nothing wrong with editing posts to make them accurate/clearer/more complete.
You might rethink the repetitive use of of multiple exclamation marks at the end of each sentence. They make it less likely that your comment will be taken seriously.
The second picture you used when referring to modern airliners and redundant systems was the cockpit of a DA-40 or Diamond Star. That's a small, single engine training aircraft that doesn't really have any redundant systems.
I once worked for a place that had a Stratus computer. It had three banks: Memory, CPU, and I/O. Each card was paired. (There were like a dozen of more cards in each bank.) If any of the cards died, say a CPU card, its partner was already doing the same processing.The computer would automatically call Stratus Corp, who would overnight you a new card. It would be on your desk in the morning. You take it to the Stratus. Open the big door. Find the card with the red light. Pull it out. Slide the new one in. In literally just a few seconds the green light comes on and you're done! (Well, close the door, too.) Zero down time. The users never know.
I find this totally amazing. Can you swap your CPU out with the machine still up and running?
Very cool. But also very expensive. I guess this was like mini-mainframe, but it's primary mission was fault tolerance. Everything was in pairs except the backplane and the tape drive.
Disclaimer: I do not work for Stratus. I do not get any compensation whatsoever for posts like this -- or from anyone, AAMAF.
How much was it ?
@@oldtwinsna8347 IDK, but I bet it was a lot! I'll see if I can find out.
but can they run club penguin
Conflux - Music and Art club penguin rewritten
Execute the Ute *F*
Execute the Ute not anymore apparently. Rip club penguin
*sigh*
Execute the Ute 🤦♂️ yes obviously it can run club penguin because that is a game and it is 100%playable as it is not closed down yet (sarcasm)
Thanks for including a shot of my DA40 Diamond Star Cockpit with G1000 'Glass Flight Deck'
"Zero downtime in credit card companies"
Tell that to Visa in Europe ;)
I love this quick style of videos you make. The density of the information is great for general understanding of very complex topics.
4:45 "I want to rent Overwatch.." shows playing TF2 in clip xD
One of the very few rewarding experiences of working at Walmart was getting to see the store's mainframe. I wasn't sure what it was at first but it's still the only of I've ever seen.
Linus's next video: Benchmarking my 3 new Mainframes
You know what really would have made this video more valuable? Contrasting mainframes from data centers, which share a lot of the same properties on more scalable (mostly) commodity hardware. :P
yup!
I thought he said "Pedobear VPN" at first. Incidentally that's a pretty fitting name for a VPN service.
Actually that’s a terrible name for a VPN service.
@Opecuted eww
Unfortunately you are right
I programmed on a mini-main-frame, it had 256 terminals. One day in the computer room the disk drive (size of a washing machine) had skidded across the room because the read/write heads were processing 10000s of transactions and moving so fast, we bolted the drive to the floor. Luckily the cable held and there was no problems.
The mainframe is your friend! The mainframe is like church, many attend, few understand!
Good presentation and explanation of Mainframe computing. Thank you for the video.
2:55 I want that "woah" as an mp3
I LOVED your video about mainframes. Well summarized. And actually, I saw one /360 working! I am a mainframe specialist, and although I am finding difficult to find another project to work on, I still love mainframes!
Define "mainframe specialist"?
If the first thing that comes to mind when you hear "Big Iron" is anything but Marty Robbins, your life is sad and I feel bad for you.
Used to do work using a mainframe, back in the 80s, at a university. Then it was Macs and PCs. I had no idea that they were still in use until I saw this video.
"what do you think of when you hear 'big iron'?"
Marty Robbins... no, uh
Fallout New Vegas...
Arizona Rangers... aw nevermind.
linus' "but even so" is very satisfying to hear idk why
"Play Overwatch on a 15 Monitor Setup"
>Guys in the Video playing TF2
>Tf2forlife
Nice video - a good friend work with a nice IBM Mainframe.
Do you ever just wipe for so long that you go 'thats good enough'?
Nope... One and Done... That's my motto.
gelatinous cube I hate that so much. Why won't it just come out
Do you ever think to yourself why dont i have a toilet that sprays my ass furiously with water?
Use water you Neanderthal! takes less than 10secs. Bring a bottle of water with you if you have to
viralarchitect LMAO!
Thanks for this Team Techquickie. I've been studying these topics for quite a while now.
Tired of having companies track your internet usage? Join our VPN and give that data to us instead!! and also pay us for taking it!!!
Bold to claim they are tracking our data, do you have any proof that they are?
+Authentically Authentic
It's really not that bold. All they have to do is throw WireShark, which is free, onto their network and have it log all the data that comes in.
whether they can or not is a moot point, wireshark/similar tools is taught in basic networking classes not to mention much more automatable methods should be easy for the engineers tunnel bear has. The question is whether or not they do it and like linus says in I think every full length sponsor spot for them, they have a top notch privacy policy that is VERY descriptive about the little data they do collect and what they do with said data, hell they don't even ask for your last name anymore because privacy
-edit and besides this way we can spite our ISPs for trying to sell our data by giving them a bunch of useless junk
+Jon Rueckert As much as in all honesty I don't actually think they collect private data, (A) collecting your last name isn't really relevant to Internet packet sniffing, and (B) what's preventing them from simply ignoring their own publicly stated policy?
i could be remembering this wrong but does a company not have to abide by its privacy policy by law in canada (which tunnel bear is based in) not to mention the really bad press
Nice vid, you really helped me understand what a mainframe is, nice work with the effects too! :)
Biggest difference is that mainframes runs one of few big instances (TB of size in RAM). For example data from credit cards around the world. This is why it must be centralized in one machine and with such a overengineered throughput. You don't want failure of your money transaction because of incorrectly updated database.
On the other hand, supercomputers usually splits problem between nodes and then just collect results. Since that all node runs independently. So computers don't really knows what how is computers around them doing.
For example. If you upload youtube video in Paris it will take some time until they move it to datacenter in west US so local viewers can access it. In case of youtube video it is not & problem, but it might be problem if your customer make one credit transaction in Paris and right after second in LA, before US datacenter is updated.
A small point of detail. I work on systems like this for a living. The Term "mainframe" is technically outmoded. New mainframe are actually designated System Z. The newest model being the Z13. So the proper term here is IBM system Z. HP equivalent (I use this term very loosely) would be the now defunct "Superdome 2"
0:03 To the town of Agua Fria rode a stranger one fine day
Indeed
My surprise was how much Random Access Memory an IBM 360 had 8 Mb in the early 1970’s . They were huge machines.
3:40 you can see that woman's eye through her hand
RSher my god you're right
RSher that means she saw the camera
Oh gosh I thought that was a Halloween mask. XP
As someone who worked with a variety of IBM and Wang Mainframes in the 1970's and 1980's I feel well qualified to answer this question. In 1974 the IBM computers were as big as a small truck with extra equipment attached that was often the size of a dozen small cars, They filled a medium warehouse sized space and required air conditioning and three phase power to operate. The hard drives had a capacity of 30 Megabyte and each unit was the size of a washing machine and the disks which were removable weighed around 40 pounds each and were the size of a small car tyre. As for the WANG equipment in 1980 it was about a quarter of the space and had lost many of the additional equipment pieces but still used hard drives of a capacity of 50Mb and about half the thickness of a small car tyre. In the IBM the total memory capacity of the computer was about 64Kb while the WANG unit was about 256KB. By comparison your smartphone is approximately 1000 times faster and 100,000 times smaller.
You forgot water cooling for the ECL CPUs and a cooling tower.
You mean the Wang VS midrange line? I had to fight tooth and nail to keep them out of our department. The parent company wanted to force all the departments to convert over to Wang and we did not want to spend the money to go through a conversion nightmare.
"i wanna run overwatch on a 50 monitor setup"
*shows tf2 on screens*
theres the comment I was looking for
TF2 is better than Overwatch
lmaooo
Spell icup
I was skeptical of this video, but this is actually a pretty good explanation. I write firmware for these things and I usually have a pretty hard time explaining it to lay people.
but can it run Minecraft on low settings?
No
Haha your cute. Of-course it can! The newer ones anyway. The mainframe runs Linux and has Java. I have put a Minecraft server on my 2004 era mainframe and it ran... but was too slow...
YES IT CAN, Just not very well... The newer machines theoretically it might be OK.
Yes.
www.reddit.com/r/Minecraft/comments/1dxul1/my_experiences_with_running_a_minecraft_server_on/
WOOSH
In the late 90s the Cleveland airport still had vintage plain text computer monitors showing flights to travelers and large dot matrix printers at gates. Y2K and 9/11 really sped up upgrades, but in truth that old stuff was probably working just fine.
"Play Overwatch".... Is busy playing TF2. Niiiicceeee
I worked on (programmed and lead software design teams) for “Big Iron” mainframes (360/370, etc) for almost 40 years. When mini computers (which didn’t last) started to come on the scene, talk was that they would “replace MFs.” I had a DEC mini-computer person call me up to understand how we solved a very complicated statistical problem on a MF. I spent about 5 minutes describing the process we used (which involved creating millions of extracts to be sorted to inflate our small amount of sample date to project the results at 4 organizational levels of the universe of data answers). When I finished I asked him, “Can you do that on a mini computer?” All I heard was silence and then he thanked me before he hung up.
I saw a picture of a T-Rex chewing up some mashed metal and the caption said, “Mainframes are back and they are Mad!” I used to say that PC’s are fast but Mainframes SCREAM!
And look how far we’ve come with “Big Iron” MFs since then. Mainframes SCREAM!
Actually more and more financial institutions either abandon zOs completely to more container based processing (microservices, DC/OS, Docker), even moving their DB2 off Mainframe and going the Oracle route or they move to Linux VMs on z/OS and run their J2EE services on that.
This process has started in the early 2000s when I was working for IBM, actually a few national banks actually moved from Mainframe to Windows in Azure! (Yeah I know crazy hey!)
The biggest problem is that native z/OS developers are hard to find and C/C++/COBOL/REXX are not really suited for the highly complex integrated solutions we have today. Business logic can become so enormously complex and distributed that you need an Enterprise Framework to help you cope with this madness.
Most big banks here in NL have a host of J2EE solutions for their enterprise software infront of their mainframe and the mainframe just does the financial transactions offered from the Enterprise software. I know for a fact because I helped build a continuous deployment solution.
Also you can look at the terminal at an airport but most airlines have a screenscraping/GUI interface over the SITA system.
" the mainframe just does the financial transactions offered from the Enterprise software."... which is what it's designed to do. If the migration process started in the early 2000s, and 96 of the world's top 100 banks and 23 of the top 25 retailers are still using them, then I'd say they realized that a rats nest of microservices wasn't up to the task of replacing their big iron.
bingo. At least those companies that want to be competitive in today's world. ; ). Actually its surprising how many are migrating to cloud/VM architecture. McDonalds, GE etc. The screen scrape thing makes me want to hurl.
Some just use 3270 emulators but not using Bisync any longer.
That green screen you see might also be IBM i (AS/400) also still alive and well since 1987. For pure transactional power, the largest IBM Power9 system running IBM i (2.74M CPW on 192 cores) now bests the largest Z15 (1.25M CPW on 190 cores) (MIPS converted to CPW). Power10 will also beat Z16 core/core.
technically, supercomputers are multiple computers, which multiple CPUs and cases. How do they get all those separate machines to act together?
Parallel computing is the term. Google it if you want, it's an interesting topic.
Thanks!
>How do they get all those separate machines to act together?
Networking.
Yeah. I am studying in Japan and I visited Japan's fastest super computer "Kei" in Kobe. The processors are made by Fujitsu and they are interconnected via their proprietary MPI system. Each processors is directly connected with 6 processors in the network in like 3 dimension like the atomic structure of a crystal. I forgot the exact name of that connection but it sounds like 3D blah blah blah.
Middlewares.
I am glad you created this video. I used to work on a pair of mainframes at a large global car manufacturer and I assisted there with the mainframes SNA connectivity amongst other things (SNA is kind of like TCP/IP but for mainframes). Mainframes are often kicked in the teeth because PC's obviously outsell them, are ubiquitous and few have true experience with them in order to make a realistic comparison to a PC. As you say, they are architecturally almost completely different and built from the ground up to work with a different kind of load to what is PC is meant for. Because of a mainframe's architecture, it can support hundreds if not thousands of users with a 1/10th of the RAM a PC would need for it to battle to cope with just one neurotic high-end user. But of course, a mainframe ain't built for making pretty pictures like a PC is. These machines live in a different but parallel dimension
.
IBM mainframe manual...nuff said
AS/400 is just enterprise software, and runs on any reasonably powerful (preferably highly multi-thread)CPU. We ran it on an older xeon chip where I work, and just replaced it 2 years ago...
The AS/400 was intended for small shops and supported job running protocols that were not compatible with the "big iron" mainframes in the Syatem 360/370/390 series.
AS/400 was IBM's midrange systems line. It was a descendant of the System/36/38/34/32/3 midrange computer line. It is now named Series i.
If you ever need a good nights sleep.
"The IBM Mainframe Manual", if printed on actual paper, small font, would not fit in the bank vault of the home office of Citibank if you cleared out all the money, lock boxes, gold bricks, and fire protection. These machines which some of you write off as uninteresting and irrelevant hold 70% of the world's data (including much of the "Big Data"), run in 97% of the "Fortune 1000" companies, and generally make your comfortable life POSSIBLE. You don't spend a dime with a credit card, book a flight, hotel, or car rental, communicate with your cell phone and/or any other small device (think laptops, tablets, etc) without what you're typing passing through, and likely being at least processed at some point by one (or more) mainframes. Whether you're interested or not, most of your life would not be even remotely as comfortable, functional, organized, easy, or safe without these machines (which of course are used to design, manufacture, distribute, and finance the cars you drive (whether purchased, leased, rented, Uber'ed or somehow otherwise shared). And, as noted, one mainframe might host (i.e. RUN) 8000 images of "real" (as defined by non-mainframers) "Intel (or other) server" machines (that all of you think of as "big machines") under a Virtual Machine (Hypervisor) running on PART of a single mainframe, along with other zOS or other operating system workloads: all this in ONE MAINFRAME. So, dismiss them as you like. Or appreciate them as a large part of every person's life in any civilized part of the planet. - "Proud Mainframer, since 1968..."
I used to operate Mainframes in the 80's, ICL ME29..... data processing for a travel company.. FUN TIMES...
Linus-"Speaking of, uh.."
Oh shit! It's time to get outta here.
worked for an airline for about a year. shit was the most reliable I worked with in a very long time.
Mainframes are frames, that are main.
Thanks - This is a great explanation of mainframes and why we may never see their end. (Lifetime job for me!)
Can you do one on Telecom
IBM Mainframes supported Graphics Terminals using software called GDDM. The API was similiar to OS/2 Presentation Manager GPI. Actually GDDM came first, so GPI was probably based on it. Mainframe graphics were not very widely used.
Before that you had Programmed Symbols, where you could make up up to 6 different fonts (9 pixels high by 16 wide) of 192 characters each. They could be single-plane (one colour) or triple plane (where you have red/green/blue planes).
Linus my professor of economics says that desktop pc's are going to die in the future because our phones are getting more powerful, what do you think about that? Personally I think he's wrong but I can't give him an specifics reason. For example if the reason is video editing, it isn't a big enough market to keep the industry going. Thanks
Felipe Prieto low power desktops will die. High power ones wont
PCs have their purpose just like phones do, and they are both different. Phones will always have the use of being the on-the-go, less powerful pieces of hardware where you can accomplish many everyday needs. PCs will always have the use of being the more potent pieces of hardware where you accomplish what simply cannot be done on a phone. The thing is, eventually we will reach a point where we cannot make our hardware any smaller to make space for power, and from that point on, we will only have to make things bigger, and as for phones, eventually they will reach their peak height and width (that which is the maximum height and width a phone can retain while still being comfortable and small enough to carry in your case, pants, purse, etc. Desktop PCs can continue to maintain their size as larger pieces of hardware with tight but large and potent parts inside. Therefore, PCs will ALWAYS be more powerful than phones, allowing them to not die out.
Felipe Prieto for consumers, yes, not laptops, but low power towers yes, maybe some will still exist for ofices and shait, but high power desktops? they will never go away, its just a fact, because of how fast they get upgrades, top of the line phones, consoles and laptops will never be faster than a top of the line desktop, simply because it takes way more time to make a complete system than just a single component to plug on a desktop, so yeah, desktops will never go away completely.
He literally made a Techquickie about this
Felipe Prieto Desktop computers serve plenty of purpose, and will continue to do so. You cannot program effectively on a phone or tablet. A physical keyboard, phone realestate and higher thermal dissipation mean desktops are the more suited form for power-hungry tasks.
Even just multi-tasking can be a pain on mobile devices.
I really enjoyed this, keep it up John and Dennis
hacking the mainframe
*Linus* Great topic. I've recently become interested in mainframe hardware. Thank you to you and your team.
"20 monitor overwatch"
*shows TF2*
i have been wanting this video for so loong!
hi
Kaitlyn Argyle hi
Kaitlyn Argyle how are you
finally a good transition into the adspot again, +1 linus
why do i have the ability to download this?
What do you mean?
Authentically Authentic right under the video on android i can, upvote or downvote the video, share, "save", or and to playlist. mostly referencing the save button.
and by clicking the vutton I have my answer, youtube red 👎👎
If you have UA-cam RED, you can save it to watch it offline later.
Oh, sorry, you replied while I was typing ;)
I started working with mainframes a few years ago and am still stumped on explaining what I do for work. I think this will actually help explaining my job a bit better. Thoug it still ends with "I get paid for keeping it alive". lol
lol
it's a good idea start a career in mainframe this days?
Second
Great content! I've been as web developer for a while and didn't even really understand what mainframes are.
Early mainframes, and I worked on some in the 70's couldn't just be turned on. You had to power it up, hit a set of switches and press execute several times to give it the instructions to read a tape strip or cassette. Then that info would let it read a 7 or 9 track tape that let it read the OS tape. Older systems would read cards.
@Richard Conte in the real old days you turned power on, then presses switches that entered a short tape or card read program . The system then read about 20 cards or a short tape that gave it a very short set of instructions. Then load a larger tape or stack of cards that held the OS then you typed xqt and tape/card identifier it would the n load the operating system. You then loaded the COBOL, FORTRAN systems, then you ran your programs such as payroll, parts ordering, etc. It was time consuming, you may only have 1,024 bytes of memory and storage was maybe 640K bytes on a drum.
@@terrydouglas5008 is it true also for the MVS operating aystem?
@@terrydouglas5008 and how do you input an information that use a different language that does not use the latin alphabet like arabic chinese etc?
@@jirehla-ab1671 it uses binary, has nothing to do with a human language
@@terrydouglas5008 let's say I want to input a person which is in arabic, so how would that be possible if your limited to ascii?
Finally a subject I understand. Did you know mainframes have never been hacked?