Its like being back at school, but fun. 10 minutes of edutainment. No filler, just what you need to know without being treat like a child or an expert. It's a YES from me
Effectively it seems like a reversion to how things used to operate, local servers handling the bulk of processing. 20 years ago, with limited internet connection, most security and manufacturing monitoring services were done at a local point with a secondary feed to an off-site cloud accessible server. Seems like it took all of this time for people to realize that local hosting was actually the better option.
I was thinking the same thing. In a way, it's similar to how virtualisation caught on yet the concept was originally used on mainframes. "What goes around comes around" indeed.
I agree, I see the Cloud as just an extension of the client/server- your end is a thin client essentially. What has made it work is ARM/RISC because it means you can off-load all the ‘heavy’ 86_64 stuff to a remote server and your end is just receiving and sending data to be crunched. And that puts pressure on bandwidth and so we get ‘EDGE’? Keeping some stuff close to hand because now you can, I noted Chris used the Latte Panda as an example and I think he’s probably right there. RISC servers seem to be restricted to a limited number of user cases but even a relatively low powered 86_64 based system can do quite a bit these days. Look at Intel’s Atom range of server chips, they are heavily feature packed and you know they could be more feature packed but Intel doesn’t want to cannibalise the product stack. The Atom was a failure in the smartphone/tablet/2in1 space but it will live on as an EDGE solution it seems. And I’m sure that AMD will be introducing their stuff rapidly as well- there are server motherboards out there. Expensive for us but cheap for enterprise. Certainly Threadripper hurts ARM because ARM servers have always been ’never mind the drawbacks, count the cores’ but AMD seems to be able to easily ship 64 cores in a cost effective manner and give you 86_64. And as you say, things like Azure/AWS sound a lot like the old time-sharing on min-computers. It’s just a natural growth as the technology has developed. It’s all before my time, and my education on the period is limited to 30 points from a OU course over 15 years’ ago, but I wonder if the modular/ distributed computing thing is new/different?
@Peter Andrijeczko Depends if you are paying or not, paying you get to protect your data- free not at all really. In my previous job we were doing highly sensitive stuff funded via the (Welsh ) government and we wouldn't be allowed to use anything that wasn't water-tight. We (I wasn't involved in the choice) went with Azure over Google.
These videos remind me of the Open University courses that would broadcast through the night on BBC2. They are very well put together, you are a great teacher.
I have been researching on edge clouds since 2014 and I commend (and approve) your work in this video. Not even many researchers have figured that "Fog" is a marketing term by CISCO to push their products. Edge, Fog, Mist, Mobile Edge, Mobile Access Edge, etc. etc. Names may be many but you hit the nail right on its head with the basic explanation of the concept. Job well done!
Thank you for adding much-needed to detail to this fascinating subject ! As everyone below is saying, it inevitably is re-inventing the wheel, which started in the late nineteenth century with the growth in use of electrical telegraphy. Operators soon had to start recording messages received, on paper tape for later onward transmission. As people below are saying, certainly during my time in the computer business ( from the late 1960s to 2000 ), centralised and localised processing regularly came in and out of favour as quickly as the changing styles of clothing worn by the operators ! The briefing that you're giving us in this video is certainly bringing us "up-to-date" with "where we're at" currently - much-needed as systems become much cleverer ! THANK YOU for this - and keep telling us !
It amazes me, how fanatic we are to give away our personal data, to get the IN thing , IOT devices will be the most intrusive items for data gathering there is, and how many apps thst you dowload for your Smart phones want access to your. pictures, contacts, files, camera, microphone etc.
@@Bithros At the moment, mostly advertising and market-research companies, at least in the 'free' world. But, and this is the important part, *that can change at any time*. Once you've given someone a copy of your house key, you have no control over who else that person gives it to. The datasets are *huge*, and can be very easily abused. People in the past have used datasets (manually tabulated, but datasets none the less) to do evil things in the past. We haven't suddenly become vastly better as a species since then either. If the Stasi, Gestapo, KGB and CIA murdered quite a lot of people based on the information they had on paper files, what do you think a despot might do with the information they could glean about you now? That's what Scott is talking about.
@@DFX2KX i don't believe that in our days people care about their data except from their bank account! If you want private data you have to pay for it if you want "free" (99% of population) you give your data! There is no other way, you can not live today without Google etc. If they want they can destroy lot's of people because of their political opinion but that happened almost all the time. Erdogan did that already. Once i chat with a Turkish guy and it was obvious that he can not say his opinion when it comes to talk about the moustache guy. But we can not do anything about that if they want they would come for us!
this is a bad idea, a break in one cable will take out a bunch of devices instead of just one. They'll rethink this when someone loses millions of dollars because their robots arent building cars one day.
Thanks, Christopher! Watching the programs you make, in particular, the ones on your AMD build, gave me the confidence to build my own AMD system. I love learning from you, please keep teaching.
Thanks for a clear explanation of Edge computing Chris. As SBC processor are getting more processing power, it makes sense to using the Edge environment. This means also that we require additional interrupt capabilities for real-time processing on the SBCs.
Process control instruments have been moving in this direction for many years with I/O that can be connected to instrumentation nearby and mounted directly on or near the equipment. The I/O communicates to the PLC via Ethernet. Now, we have highly accurate devices with built in diagnostics that connect to the PLC over Ethernet. Previously, multiple wires for analog and discrete signals would have been required to send the information for mass flow, density, temerature, flow total and device health from a coriolis meter and now, using Ethernet IP or Modbus TCP, all of that information and more is sent directly to the PLC, HMI, or process historian. I think the key to acceptance of SBC devices on the plant/factory floor will be availability of "open" standards like OPC, EthernetIP and ModbusTCP. Thanks for another thought provoking video.
Back in 2010 you made a video called “3 ways to cloud compute. “. It appears that you may need to add a 4th way. Today’s video is a great update. It should be mandatory reading for anyone going into university.
It's a never-ending debate between one large central computing powerhouse and distributed, smaller computing powercells. Look at how mainframes, terminals and home computers were used in the 20. century. It's the exact same think, except with the Internet thrown in for good measure.
We used to call this "distributed" computing as made possible by local area networks. Calling it "edge" and "mesh" is cute, but it really is just recycled tech.
Excellent video! The information is relayed clearly and concisely. Before watching this I never knew anything about Edge Computing. Now I do. Result! Thank you
Another great video of just the right length. No waffle, no banter, just the information you need to understand the concept. It makes me wonder, is this process cyclic? 50 years ago compute power was expensive so processing was centralised with mainframes and dumb terminals. Then PCs change the landscape with cheap processing and we de-centralised. Then we upped the amount of processing we needed to handle the vast amounts of data we are now collecting so the Cloud was born and we centralised again. Now we've got dedicated devices like VPUs and pre-trained neural nets that can offload the processing and a limited resource of bandwidth, so we're decentralising again. I wonder if we "fix" the bandwidth problem will we centralise again ...😀
Thank you Christopher for this video. I'm new in the sw consultent business and we are going to work with these things. What you have done is to very pedagogically explaining how every thing is related.
Your explanation of cloud and edge computing reminds me of the old client server model that evolved into cooperative computing where the client PC would process the data before transmitting to the server. Lots of new terminology.
An interesting and well explained video - thanks. It shows how the internet added the ability to draw on more powerful resources, not available locally, to process data. But, as the quantity of date increased that travels over the internet, the internet's attributes became the limiting factor; as does the number of requests. The availability of powerful and relatively cheap single board computers has allowed processing of date at the edge, where is is created, of a central location, being the cloud. Thus, there is an opportunity for individuals to have control and ownership of their own data. Rather than hand it over to large corporations and lose control of it. Moreover, in edge computing there is opportunity to learn about AI in good detail, without a need to work for a large corporation to gain this knowledge. As always we live in interesting times, as change appears to be the only certainty.
Interesting how the traditional mainframe/mini/terminal arrangement changed to 'smart' terminals (AKA PCs) and now the old mainframes are the new clouds. of course, in those days, edge computing was done with a deck of punched cards. ;). Times have changed... need to upgrade my 029 Keypunch one of these days! Heh!
Brilliant explanation of the this technology. Thanks for taking the time to explain in such detail and with common language. Really enjoy your channel. Keep up the great work!
Thanks Chris. Some places have poor internet access (like me 2 megabits down, 100 kilobits up on a good day) so doing the processing locally is very important.
I have been using "fog computing" to describe the deviation between the theoretical cost/performance improvement of cloud computing with time and the rising cost of the actual billing with no increase in productivity. Management is cutting cost by moving to the cloud then giving themselves a huge bonus. Once they are in the cloud IT becomes an incontrollable mission critical expense because they got rid of everyone capable of moving to another cloud vendor. Fog computing is the extra margin being paid on cloud computing due to the poor strategic movement of mission critical functions to the cloud without retaining the expertise to manage and move the business. If you can't move your cloud to another vendor there is no reason to believe that you will maintain the low commodity price used to justify moving to the cloud. I am biased, but I think my definition of "Fog Computing" is most useful.
"Edge = Skynet eyes". I like it. :) And basically true! Lots -- probably millons and millions -- of little boards keeping an eye on things with thier neural nets, reporting back when they come across something of note . . .
Edge computing is a logical step. Simplest way to unburden the internet. NN, ML, huge amount of data..., but as well powerful SBCs, faster machines and greater storage makes edge computing an expected step. But more important, much more possibilities to experiment with, on individual level, affordable technology with power. Citius, altius, fortius. Kind of'reversing', before cloud computing era. 🤔 Excellent video! 👌🏻✨ Cheerio Chris! 👋🏻
Thank you for the enlightening and well made report, Christopher. At 2:54 you describe the case of the actual processing taking place at the edge and not as before in the cloud. Then you state, that only the result of that computation is sent back, improving latency and efficiency of the system. This is an important point! If we account for locality then in a real life scenario, the result is needed in the physical place at the factory and not in the cloud. So a blasphemous thought could be that the transfer to the cloud is a detour, which in turn puts the cloud computing metaphor in question. I've always suspected that;)
You are right, sometimes the result will only be needed locally. But the neural net will still need to be trained (and retrained over time), which is likely to take place on cloud servers. But the balance of how much takes place locally and centrally will always be in flux, depending on the particular situation.
@@ExplainingComputers Exactly that thought, how these responsibilities in manufacturing or IT-aided processes can shift in almost a fuzzy way brought me there.
I remember back in the 90's when Scott McNealy said "the network would be the computer", (or something very similar) and I guess now that everybody has figured out it costs a lot of money to build and maintain the centralized "cloud", those "cloud owners" are trying to force that computing resource (and costs) back on the users, (or should they now be called "edge users"?) to lower their own cloud costs. Sound a lot to me that we have come full circle.
You forgot to look at the negative side of Edge computing. Security threats of having more readily exposed data and access of local networks endpoints, Less energy efficiency of more hardware operations and lost computing cycles, and impact on environment and mineral resources of having to manufacture and deploy more SBC and local servers hardware, and last but not least the cost of maintenance and support required to keep the Edge units working.
good explanation. edge computing seems a bit like an arbitrary term. it's like people suddenly realized "Hey we can't move everything into the cloud. We have to move some of it back!"
Another great video Chris! It would be great to see a future video explaining the technical security challenges of 5G hardware. I think understanding the tech can be apolitical, in the same way that we can understand general challenges of internet security without knowing specific bad actors. You have a knack of making it simple to understand!
Wyze Cameras Use Xnore Ai and the algorithm is trained then placed on the firmware therefore processed on the camera then transmitted once a particular value ( detects a person) is met, this is interesting, as I was just learning more about this myself to help resolve some resource/network bottlenecks, Excellent Video!
Very interesting. This shows that it's important that SBCs continue to improve, so that they can process data on their own rather than just be the modern equivalent of the dumb terminal.
Thanks for an excellent video about Edge computing. One the key reasons, I enjoy your channel is that you keep me updated on new and emerging technologies (and implementation of the technologies in new innovative ways). Please keep the excellent work!
after last week's video I was thinking how a modern office could run quite easily with e.g.. 4 raspberry pi s, one as a NAS and three workstations, and it made me think how we seem to taking a step backwards to actually go a step forward. Why backwards? well it resembles when we had computer that were not very powerful and would have either network server and or a large hard drive (yes those ones you would screw the HD to) that the work stations connected to. Then with progress etc. the office server environment changed especially in the small office, PC s with much more power and storage capacity became available. Now watching today's video the ping pong thought of where the procesing and storing happens pops back into mind.🤔
personally i dont like the cloud , or keeping my info on other peoples computers because thats all the ' cloud ' is , i love openhab and keeping my info to myself
People forget that mainframes were shunned because of this and that's what powered the PC revolution. We are regressing by moving back to a mainframe mentality. Edge basically proves that there's little need to use cloud services and relinquish your data to a corporate monolith.
Cloud computing is awesome, I think when people think about “the cloud” they are thinking about iCloud or google drive. You can make your own cloud computing network if you’re worried about your data 🤔 but whats “secure” in this day and age right lol
Some people are commenting as if cloud is something bad and services are reverting back to "good old days" where local machines did all the processing. That is a fundamental misunderstanding of cloud and edge computing paradigm. Some services cannot be processed at the edge and more computationally intensive operations are still tasked to the cloud. Modern SBCs are powerful enough to lend a hand to the cloud. This seamless sharing of resources which has ushered in a host of new applications and services has never existed before.
Exactly. Edge computing does not replace cloud computing -- edge devices are all cloud linked, and do some data (pre)processing local to expand cloud uses as you describe. Somebody below called in spine computing, which is a great analogy. The spine does some basic processing, taking some of the load off the central brain.
Very good video! I've never heard of this before, but it strikes me that 'Edge computing' will be necessary for self driving, autonomous cars in the future.
Very similar in principle to what used to be the old network based Microsoft Terminal server.. The end clients were able to do a small amount of processing, with the bulk done by the main Terminal Server. That was back in 1999.
One thing you didn't mention is that edge computing potentially gives the users back some control over their data and the devices processing it. No longer would all data be sent into an ominous "cloud" (which is a marketing term for "somebody else's computer") but be able to be processed under the oversight of the user creating it. Other than that gripe, that was an excellent overview!
Thanks for your awesome educational videos. I love how you stay on topic and leave conjecture and mockery out of your content. Though these things have their place (such as comedy, or satirical jabs to ultimately aid developers), it is too often over-done. Perhaps we shall see your videos sold on Amazon video, or for sale in educational libraries.
I think edge computing could be very helpful for individuals, and not just for enterprise users. I believe it was in the comments of one of your videos - or perhaps the video itself - that someone talked about the notion of a databox that would use a pretrained neural network to take in all the data from devices in your home and either make inferences by itself or convert that data to something more acceptable to send to a larger datacentre. The reasoning behind it is that the company that makes your smart teapot doesn't need to know the Unix timestamp of when you put the kettle on if it only wants to know how many cups of tea you have in a day, so the databox would do all the logging and send a revised average of cups of tea per day to the datacentre at the end of each week. The cool thing about it is that your entire house could automate a lot of things while being offline; e.g. your smart shower would tell the databox that its just been turned on, and the databox would infer that you don't want to step out of the warm shower into a cold room, so it turns the heating up a little. The databox also uses information from your smart security system to know what your usual route is after you've taken a shower, and draws your smart blinds so people can't see you walking to the bedroom, and all of that without ever touching a computer that's not in your house, or sending data to a device that doesn't need to know.
In your mentioned factory example. It should not send the defective information to the cloud. It should send the wrong inferred (decision) information back to cloud in order it to refine the AI program which will then send back to factory for the replacement of less effective one.
What came to my mind was my highschool network server. It connected all the computers together so you can log onto your account on any computer in the school.
Computing like fashion always goes in cycles, moving from centralising to decentralising; Unix (centralised), Windows clients (decentralised), Cloud (centralised), and now Edge (decentralised), give it another 10 years and everything will be heading back to the centre again.....
Thanks for another excellent tutorial. I could never work out how training your Raspberry Pi to tell the difference between an orange and a banana was going to change the world.
You won't get an increase in intelligence but you will get an increase in knowledge. Personally, I struggle with both these days approaching my seventh decade. :/
@@retiree1033 Well me being an Elder Geek, who has just reached my sixth decade on this planet.. I have the intelligence and ability to acquire and apply knowledge.. Peace!
Another excellent video.. Just enough to whet the appetite. It is funny, I have been in technology my entire life and on the edge of it for many years (retired) and we always joked centralize/decentralize /centralize/decentralize. So this will happen for another era and then we determine oops too much to manage decentralized and since we have even more powerful cloud computers and faster internet lets centralize again. LOL! I am sixty-five and still as interested today as I was back then regarding technology. Thanks again.
Watching computers evolve the way organic biology did is amazing. This is just like a video I watched about "robot skin", how there's too many points of data to track even for a human mind so a lot of the sensing data is computed by the nerves before it sends a signal, basically just significant changes are sent to your brain, and this is the exact same thing for silicon based intelligence. Watching the entire planet gain sentience via computing is wild, it's a crazy time to be alive to watch this all happening so fast right in front of us. It took people billions of years to reach this point, and computers will surpass it in a hundred. What massive and exponentially more potent use of matter, it's almost insane to think about. This of course, is completely leaving out the ramifications, the philosophy behind this. It's just a fact, computers are becoming more organic but with the ability to evolve in seconds, not millennia. It's organic life, but with the "edit" mode unlocked.
Ah, you are a mind reader Bill. I included mesh networks here because I have plans for a video in which I build a mesh network using several Raspberry Pis. :)
Thank you for the Sunday Lesson as always Chris! Sorry to sound facetious but doesn't Edge Computing sound rather like what we had before the cloud? :P
This really seems to me like old 3 tier client server setups were the clients would communicate with powerful second tier server and then push back up to the main server when required. I find amazing how how we do seem to be going backward ie originally central large main frames with dul terminals then we went to powerful PCs and the 3 tiers client server then back to centralised cloud and now back to client server setups. I guess it all helps hardware providers sell more hardware in the and like retailers change the fashion to sell more clothes and how fashion goes in circles
so i'm 3 minutes in, but i think i already have a grasp of it. Edge computing in things like Facial recognition software helps in.. well.. speed and efficiency, since Edge computing is as the name suggests, as close to the source as possible. Very interesting stuff indeed.
Its like being back at school, but fun. 10 minutes of edutainment. No filler, just what you need to know without being treat like a child or an expert. It's a YES from me
Can you press the Golden Buzzer please?
Edutainment. I see what you did there.
Eeenà00ý
@@boredchicken5755 qqq9p9i onilqp
Effectively it seems like a reversion to how things used to operate, local servers handling the bulk of processing.
20 years ago, with limited internet connection, most security and manufacturing monitoring services were done at a local point with a secondary feed to an off-site cloud accessible server.
Seems like it took all of this time for people to realize that local hosting was actually the better option.
Agreed; what comes around goes around, yada yada!
I was thinking the same thing. In a way, it's similar to how virtualisation caught on yet the concept was originally used on mainframes. "What goes around comes around" indeed.
Adobe: "D'oh!"
I agree, I see the Cloud as just an extension of the client/server- your end is a thin client essentially. What has made it work is ARM/RISC because it means you can off-load all the ‘heavy’ 86_64 stuff to a remote server and your end is just receiving and sending data to be crunched. And that puts pressure on bandwidth and so we get ‘EDGE’?
Keeping some stuff close to hand because now you can, I noted Chris used the Latte Panda as an example and I think he’s probably right there. RISC servers seem to be restricted to a limited number of user cases but even a relatively low powered 86_64 based system can do quite a bit these days. Look at Intel’s Atom range of server chips, they are heavily feature packed and you know they could be more feature packed but Intel doesn’t want to cannibalise the product stack. The Atom was a failure in the smartphone/tablet/2in1 space but it will live on as an EDGE solution it seems. And I’m sure that AMD will be introducing their stuff rapidly as well- there are server motherboards out there. Expensive for us but cheap for enterprise. Certainly Threadripper hurts ARM because ARM servers have always been ’never mind the drawbacks, count the cores’ but AMD seems to be able to easily ship 64 cores in a cost effective manner and give you 86_64.
And as you say, things like Azure/AWS sound a lot like the old time-sharing on min-computers. It’s just a natural growth as the technology has developed. It’s all before my time, and my education on the period is limited to 30 points from a OU course over 15 years’ ago, but I wonder if the modular/ distributed computing thing is new/different?
@Peter Andrijeczko Depends if you are paying or not, paying you get to protect your data- free not at all really. In my previous job we were doing highly sensitive stuff funded via the (Welsh ) government and we wouldn't be allowed to use anything that wasn't water-tight. We (I wasn't involved in the choice) went with Azure over Google.
These videos remind me of the Open University courses that would broadcast through the night on BBC2. They are very well put together, you are a great teacher.
I have been watching this channel for 10 years now, and this guy hasn't changed a bit! I love it!
Check his attic for an "aging" portrait!!!🤔🤔🤔😀😀😀😀
I have been researching on edge clouds since 2014 and I commend (and approve) your work in this video. Not even many researchers have figured that "Fog" is a marketing term by CISCO to push their products. Edge, Fog, Mist, Mobile Edge, Mobile Access Edge, etc. etc. Names may be many but you hit the nail right on its head with the basic explanation of the concept. Job well done!
Thanks for your positive feedback. Appreciated. :)
Thank you for adding much-needed to detail to this fascinating subject ! As everyone below is saying, it inevitably is re-inventing the wheel, which started in the late nineteenth century with the growth in use of electrical telegraphy. Operators soon had to start recording messages received, on paper tape for later onward transmission. As people below are saying, certainly during my time in the computer business ( from the late 1960s to 2000 ), centralised and localised processing regularly came in and out of favour as quickly as the changing styles of clothing worn by the operators ! The briefing that you're giving us in this video is certainly bringing us "up-to-date" with "where we're at" currently - much-needed as systems become much cleverer ! THANK YOU for this - and keep telling us !
Well done sir.
Getting up at 6am to start my Sunday with your videos isn't fun, but definitely worth it, thanks Chris.
Thanks Ted. In which country is it 6:00am right now?
Its 9am in Florida right now
@@ExplainingComputers Explaining Computers comes out 6am Sunday morning in California, US of A
11:00am in Argentina
It amazes me, how fanatic we are to give away our personal data, to get the IN thing , IOT devices will be the most intrusive items for data gathering there is, and how many apps thst you dowload for your Smart phones want access to your. pictures, contacts, files, camera, microphone etc.
I too am amazed by this. But all the evidence is that most people are "willing" to share in return for perceived value-added services.
Who cares about your and our personal data etc. Even if we try to make people download it probably nobody want them.
@@ExplainingComputers Not me I deny any bloated apps, or downloaded apps if they want acces to my personal data or camera, microphone etc.
@@Bithros At the moment, mostly advertising and market-research companies, at least in the 'free' world.
But, and this is the important part, *that can change at any time*. Once you've given someone a copy of your house key, you have no control over who else that person gives it to.
The datasets are *huge*, and can be very easily abused. People in the past have used datasets (manually tabulated, but datasets none the less) to do evil things in the past. We haven't suddenly become vastly better as a species since then either. If the Stasi, Gestapo, KGB and CIA murdered quite a lot of people based on the information they had on paper files, what do you think a despot might do with the information they could glean about you now?
That's what Scott is talking about.
@@DFX2KX i don't believe that in our days people care about their data except from their bank account! If you want private data you have to pay for it if you want "free" (99% of population) you give your data! There is no other way, you can not live today without Google etc. If they want they can destroy lot's of people because of their political opinion but that happened almost all the time. Erdogan did that already. Once i chat with a Turkish guy and it was obvious that he can not say his opinion when it comes to talk about the moustache guy. But we can not do anything about that if they want they would come for us!
Fog is low level clouds... Got it.
You put it better than I did!
Cloud is also that guy who wields that hugeass sword lol
True in two ways
bruh
this is a bad idea, a break in one cable will take out a bunch of devices instead of just one. They'll rethink this when someone loses millions of dollars because their robots arent building cars one day.
Thanks, Christopher! Watching the programs you make, in particular, the ones on your AMD build, gave me the confidence to build my own AMD system. I love learning from you, please keep teaching.
Great to hear that you did your own build! :)
Mainframes, then PCs, then the cloud and now edge computing. Back and forth it goes!
...as the pendulum does. with a massive increase in gravity with each swing, though.
Thanks for another exciting show. Edge computing is much needed when speed and response time is important.
Thanks for a clear explanation of Edge computing Chris. As SBC processor are getting more processing power, it makes sense to using the Edge environment. This means also that we require additional interrupt capabilities for real-time processing on the SBCs.
Process control instruments have been moving in this direction for many years with I/O that can be connected to instrumentation nearby and mounted directly on or near the equipment. The I/O communicates to the PLC via Ethernet.
Now, we have highly accurate devices with built in diagnostics that connect to the PLC over Ethernet. Previously, multiple wires for analog and discrete signals would have been required to send the information for mass flow, density, temerature, flow total and device health from a coriolis meter and now, using Ethernet IP or Modbus TCP, all of that information and more is sent directly to the PLC, HMI, or process historian.
I think the key to acceptance of SBC devices on the plant/factory floor will be availability of "open" standards like OPC, EthernetIP and ModbusTCP.
Thanks for another thought provoking video.
Back in 2010 you made a video called “3 ways to cloud compute. “. It appears that you may need to add a 4th way.
Today’s video is a great update. It should be mandatory reading for anyone going into university.
Hi Dale. Things indeed move on. :)
So Edge Computing is reversing/ take computing power out of the cloud but without admitting it..
Yes, although some would say it was adding even more power to the cloud! :)
flexairz “without admitting it” ?
It's a never-ending debate between one large central computing powerhouse and distributed, smaller computing powercells. Look at how mainframes, terminals and home computers were used in the 20. century. It's the exact same think, except with the Internet thrown in for good measure.
If they admitted they were just going back to the old way of doing things, they could not say it was "New".
We used to call this "distributed" computing as made possible by local area networks. Calling it "edge" and "mesh" is cute, but it really is just recycled tech.
I love every video, I was a CS student and now is a software engineer, thank you Mr. Chris.
You're very welcome!
Another great video. This channel is truly educational, and I look forward to it every week.
A few years ago we were all rushing to the cloud. Now the edge is 'the thing' and we're all rushing back there. Whew! what a roller coaster ride.
This concept of Edge Computing is very interesting, I’m going to research a bit more about this.
Oh and Happy Sunday ExplainingComputers!
Thanks John. You may find the links in the video description - to IBM, the Linux Foundation and Cisco edge computing pages -- be be useful.
It's like CDN networks for computing not just static content
Excellent video! The information is relayed clearly and concisely. Before watching this I never knew anything about Edge Computing. Now I do. Result! Thank you
interesting switchback, to see compute move away from the cloud to augment efficiency, you explain things well, thanks!!
Proud Nottingam Business School alumni. I remember your lectures on cloud computing eons before it was mainstream.
Thanks for this. Great to hear from you. Those were the days! I left the university in late 2015, although I have taught there since (and at Trent!).
Another great video of just the right length. No waffle, no banter, just the information you need to understand the concept.
It makes me wonder, is this process cyclic? 50 years ago compute power was expensive so processing was centralised with mainframes and dumb terminals. Then PCs change the landscape with cheap processing and we de-centralised. Then we upped the amount of processing we needed to handle the vast amounts of data we are now collecting so the Cloud was born and we centralised again. Now we've got dedicated devices like VPUs and pre-trained neural nets that can offload the processing and a limited resource of bandwidth, so we're decentralising again. I wonder if we "fix" the bandwidth problem will we centralise again ...😀
Excellent video. I'm glad everything is explained very clearly. Thank you!
Thank you Christopher for this video. I'm new in the sw consultent business and we are going to work with these things. What you have done is to very pedagogically explaining how every thing is related.
One of the most professional and informative UA-cam channels.
Thanks.
I just shared this with my team and manager to breach the informational gap. Great work, thank you soo much!
Great to hear.
So, are smart phones not basically just wireless edge computers?
Now that is a great thought. :)
Yes. No additional hardware needed. Most of our devices should be capable of running trained models and report back if the model failed
great, So my cell phone will report when I'm slacking off instead of working. :P
theonecalledstein of course......... bUt WhO iS iT rEpOrTiNg ToO??!!!???
@theonecalledstein If your cell phone is in your hand when the boss walks by, you bet it will.
AI with Cloud computing simultaneously is big industry changer.
Agreed!
Eyy it's a new explaining computers video on my birthday! Thanks chris!
Happy Birthday! :)
@@ExplainingComputers thank you!
Your explanation of cloud and edge computing reminds me of the old client server model that evolved into cooperative computing where the client PC would process the data before transmitting to the server. Lots of new terminology.
An interesting and well explained video - thanks. It shows how the internet added the ability to draw on more powerful resources, not available locally, to process data. But, as the quantity of date increased that travels over the internet, the internet's attributes became the limiting factor; as does the number of requests. The availability of powerful and relatively cheap single board computers has allowed processing of date at the edge, where is is created, of a central location, being the cloud. Thus, there is an opportunity for individuals to have control and ownership of their own data. Rather than hand it over to large corporations and lose control of it. Moreover, in edge computing there is opportunity to learn about AI in good detail, without a need to work for a large corporation to gain this knowledge. As always we live in interesting times, as change appears to be the only certainty.
Again you manage to explain the subject so all can understand. I tip my hat to you Sir
Probably the best explanation on UA-cam. Subscribed.
Welcome aboard!
Interesting how the traditional mainframe/mini/terminal arrangement changed to 'smart' terminals (AKA PCs) and now the old mainframes are the new clouds. of course, in those days, edge computing was done with a deck of punched cards. ;). Times have changed... need to upgrade my 029 Keypunch one of these days! Heh!
Brilliant explanation of the this technology. Thanks for taking the time to explain in such detail and with common language. Really enjoy your channel. Keep up the great work!
Thanks Chris. Some places have poor internet access (like me 2 megabits down, 100 kilobits up on a good day) so doing the processing locally is very important.
This was another awesome theme, like usual. I enjoyed very much and not just ones. Good work!
I always learn something from your videos. Thanks for quality content and keep 'em coming!
Your alien voice came out at the end when you said "very soon". I knew there was something different about you. Nanu nanu!
Excellent that you've finally covered an Intel Compute Stick! :)
I have been using "fog computing" to describe the deviation between the theoretical cost/performance improvement of cloud computing with time and the rising cost of the actual billing with no increase in productivity. Management is cutting cost by moving to the cloud then giving themselves a huge bonus. Once they are in the cloud IT becomes an incontrollable mission critical expense because they got rid of everyone capable of moving to another cloud vendor. Fog computing is the extra margin being paid on cloud computing due to the poor strategic movement of mission critical functions to the cloud without retaining the expertise to manage and move the business. If you can't move your cloud to another vendor there is no reason to believe that you will maintain the low commodity price used to justify moving to the cloud.
I am biased, but I think my definition of "Fog Computing" is most useful.
Sir, you are amazing you explain in a very simple way so every one can understand what you are saying.
Love from Pakistan
Thanks.
Great content as usual, allways look forward to your videos even though I have to watch more than once to get a handle on some of the subjects.
So, those edge computers will be the eyes of Skynet! How devious!
"Edge = Skynet eyes". I like it. :) And basically true! Lots -- probably millons and millions -- of little boards keeping an eye on things with thier neural nets, reporting back when they come across something of note . . .
It's obvious that computers can be related to eachother and can take notice what happens in the envoriment, thanks for showing and kind regards.
Edge computing is a logical step.
Simplest way to unburden the internet.
NN, ML, huge amount of data..., but as well powerful SBCs, faster machines and greater storage makes edge computing an expected step.
But more important, much more possibilities to experiment with, on individual level, affordable technology with power.
Citius, altius, fortius.
Kind of'reversing', before cloud computing era. 🤔
Excellent video! 👌🏻✨
Cheerio Chris! 👋🏻
Hi Elvira! Sorry I did not see you comment earlier. IU hope that life is treating you well this week. :)
ExplainingComputers
Experimenting and testing a new Linux distro is always a good starting point of a good week. 👌🏻😁
Looking forward to Sunday!
Thank you for the enlightening and well made report, Christopher.
At 2:54 you describe the case of the actual processing taking place at the edge and not as before in the cloud. Then you state, that only the result of that computation is sent back, improving latency and efficiency of the system. This is an important point! If we account for locality then in a real life scenario, the result is needed in the physical place at the factory and not in the cloud.
So a blasphemous thought could be that the transfer to the cloud is a detour, which in turn puts the cloud computing metaphor in question.
I've always suspected that;)
You are right, sometimes the result will only be needed locally. But the neural net will still need to be trained (and retrained over time), which is likely to take place on cloud servers. But the balance of how much takes place locally and centrally will always be in flux, depending on the particular situation.
@@ExplainingComputers Exactly that thought, how these responsibilities in manufacturing or IT-aided processes can shift in almost a fuzzy way brought me there.
I remember back in the 90's when Scott McNealy said "the network would be the computer", (or something very similar) and I guess now that everybody has figured out it costs a lot of money to build and maintain the centralized "cloud", those "cloud owners" are trying to force that computing resource (and costs) back on the users, (or should they now be called "edge users"?) to lower their own cloud costs. Sound a lot to me that we have come full circle.
It also means that data collection can be kept local, which is good for data privacy and GDPR. e.g. reduces legal paperwork.
You forgot to look at the negative side of Edge computing. Security threats of having more readily exposed data and access of local networks endpoints, Less energy efficiency of more hardware operations and lost computing cycles, and impact on environment and mineral resources of having to manufacture and deploy more SBC and local servers hardware, and last but not least the cost of maintenance and support required to keep the Edge units working.
All true!
Cloud computing is the equivalent of public transportation
Clear explanation and great content as always . Greetings from London!!!
Sunday morning. Fall weather. Happy days and a EC video? Tis a good day.
good explanation. edge computing seems a bit like an arbitrary term. it's like people suddenly realized "Hey we can't move everything into the cloud. We have to move some of it back!"
Another great video Chris!
It would be great to see a future video explaining the technical security challenges of 5G hardware. I think understanding the tech can be apolitical, in the same way that we can understand general challenges of internet security without knowing specific bad actors. You have a knack of making it simple to understand!
Nice video idea. Noted!
Congratulations on passing the 512K subscribers mark! :)
Thanks. 512K does seem significant for this channel! Roll on 640K. :)
ExplainingComputers.com is always an education for me. Thank you Christopher. Just can't wait for the talking door knobs ;0)
This will be very helpful to business models that pay for cloud processing by the amount of data processed on a cloud service
Wyze Cameras Use Xnore Ai and the algorithm is trained then placed on the firmware therefore processed on the camera then transmitted once a particular value ( detects a person) is met, this is interesting, as I was just learning more about this myself to help resolve some resource/network bottlenecks, Excellent Video!
Very interesting. This shows that it's important that SBCs continue to improve, so that they can process data on their own rather than just be the modern equivalent of the dumb terminal.
Exactly. This is why I made an edge computing video, to highlight the push in industry for more and more powerful SBCs.
Thanks for an excellent video about Edge computing. One the key reasons, I enjoy your channel is that you keep me updated on new and emerging technologies (and implementation of the technologies in new innovative ways). Please keep the excellent work!
after last week's video I was thinking how a modern office could run quite easily with e.g.. 4 raspberry pi s, one as a NAS and three workstations, and it made me think how we seem to taking a step backwards to actually go a step forward. Why backwards? well it resembles when we had computer that were not very powerful and would have either network server and or a large hard drive (yes those ones you would screw the HD to) that the work stations connected to. Then with progress etc. the office server environment changed especially in the small office, PC s with much more power and storage capacity became available. Now watching today's video the ping pong thought of where the procesing and storing happens pops back into mind.🤔
I'm taking a step back even further and running CP/M on 8 bit Z80 computers. Thats what people did back in the 80's.
personally i dont like the cloud , or keeping my info on other peoples computers because thats all the ' cloud ' is , i love openhab and keeping my info to myself
People forget that mainframes were shunned because of this and that's what powered the PC revolution. We are regressing by moving back to a mainframe mentality. Edge basically proves that there's little need to use cloud services and relinquish your data to a corporate monolith.
How you getting on with eMail and You Tube?
Cloud computing is awesome, I think when people think about “the cloud” they are thinking about iCloud or google drive. You can make your own cloud computing network if you’re worried about your data 🤔 but whats “secure” in this day and age right lol
You dont like the cloud? Never play final fantasy 7... lol
Unfortunately, keeping data in your home NAS has bigger possibilities of being hacked compared with keeping data in iCloud / google cloud
Some people are commenting as if cloud is something bad and services are reverting back to "good old days" where local machines did all the processing. That is a fundamental misunderstanding of cloud and edge computing paradigm. Some services cannot be processed at the edge and more computationally intensive operations are still tasked to the cloud. Modern SBCs are powerful enough to lend a hand to the cloud. This seamless sharing of resources which has ushered in a host of new applications and services has never existed before.
Exactly. Edge computing does not replace cloud computing -- edge devices are all cloud linked, and do some data (pre)processing local to expand cloud uses as you describe. Somebody below called in spine computing, which is a great analogy. The spine does some basic processing, taking some of the load off the central brain.
Thanks Chris for a very informative and well done video.
excellent, excellent video Chris. thank you!
This is amazing simple and to the point.. i owe you my thanks❤️
No problem 😊
Very good video! I've never heard of this before, but it strikes me that 'Edge computing' will be necessary for self driving, autonomous cars in the future.
Correct! Tesla is doing this right now.
Very similar in principle to what used to be the old network based Microsoft Terminal server.. The end clients were able to do a small amount of processing, with the bulk done by the main Terminal Server. That was back in 1999.
Informative and well presented. Thanks from Orlando
Thanks my friend.
Wow back to distributed processing. That's new..
IN OTHER WORDS, taking a STEP BACK from the hype of CLOUD COMPUTING taking the world over..
Great summary for the topic, really helped culminate the topic for me. Thanks for another video!
One thing you didn't mention is that edge computing potentially gives the users back some control over their data and the devices processing it. No longer would all data be sent into an ominous "cloud" (which is a marketing term for "somebody else's computer") but be able to be processed under the oversight of the user creating it.
Other than that gripe, that was an excellent overview!
Good point. Although it also means that data on users may end up being spread across far many more devices, which is not necessarily a good thing. :)
Thanks for your awesome educational videos. I love how you stay on topic and leave conjecture and mockery out of your content. Though these things have their place (such as comedy, or satirical jabs to ultimately aid developers), it is too often over-done. Perhaps we shall see your videos sold on Amazon video, or for sale in educational libraries.
I think edge computing could be very helpful for individuals, and not just for enterprise users. I believe it was in the comments of one of your videos - or perhaps the video itself - that someone talked about the notion of a databox that would use a pretrained neural network to take in all the data from devices in your home and either make inferences by itself or convert that data to something more acceptable to send to a larger datacentre. The reasoning behind it is that the company that makes your smart teapot doesn't need to know the Unix timestamp of when you put the kettle on if it only wants to know how many cups of tea you have in a day, so the databox would do all the logging and send a revised average of cups of tea per day to the datacentre at the end of each week.
The cool thing about it is that your entire house could automate a lot of things while being offline; e.g. your smart shower would tell the databox that its just been turned on, and the databox would infer that you don't want to step out of the warm shower into a cold room, so it turns the heating up a little. The databox also uses information from your smart security system to know what your usual route is after you've taken a shower, and draws your smart blinds so people can't see you walking to the bedroom, and all of that without ever touching a computer that's not in your house, or sending data to a device that doesn't need to know.
In your mentioned factory example. It should not send the defective information to the cloud. It should send the wrong inferred (decision) information back to cloud in order it to refine the AI program which will then send back to factory for the replacement of less effective one.
Interesting SBC. Nice heat sink too.
Very informative indeed chris
Great vid, gave me a good overview of the concept!
Wow. I have been waiting for this video. i love it. Thanks
Love this kind of videos! Very educational. Keep up the good work
What came to my mind was my highschool network server. It connected all the computers together so you can log onto your account on any computer in the school.
Computing like fashion always goes in cycles, moving from centralising to decentralising; Unix (centralised), Windows clients (decentralised), Cloud (centralised), and now Edge (decentralised), give it another 10 years and everything will be heading back to the centre again.....
Thanks for the video, very clarifying.
Thanks for another excellent tutorial. I could never work out how training your Raspberry Pi to tell the difference between an orange and a banana was going to change the world.
As soon as I get here, I press thumbs up. Why? Because I know
for sure my intelligence will be increased at the end..
Thank you!
You won't get an increase in intelligence but you will get an increase in knowledge. Personally, I struggle with both these days approaching my seventh decade. :/
@@retiree1033 Well me being an Elder Geek, who has just reached my sixth decade on this planet.. I have the intelligence and ability to acquire and apply knowledge.. Peace!
Another excellent video.. Just enough to whet the appetite. It is funny, I have been in technology my entire life and on the edge of it for many years (retired) and we always joked centralize/decentralize /centralize/decentralize. So this will happen for another era and then we determine oops too much to manage decentralized and since we have even more powerful cloud computers and faster internet lets centralize again. LOL! I am sixty-five and still as interested today as I was back then regarding technology. Thanks again.
Watching computers evolve the way organic biology did is amazing. This is just like a video I watched about "robot skin", how there's too many points of data to track even for a human mind so a lot of the sensing data is computed by the nerves before it sends a signal, basically just significant changes are sent to your brain, and this is the exact same thing for silicon based intelligence. Watching the entire planet gain sentience via computing is wild, it's a crazy time to be alive to watch this all happening so fast right in front of us.
It took people billions of years to reach this point, and computers will surpass it in a hundred. What massive and exponentially more potent use of matter, it's almost insane to think about.
This of course, is completely leaving out the ramifications, the philosophy behind this. It's just a fact, computers are becoming more organic but with the ability to evolve in seconds, not millennia. It's organic life, but with the "edit" mode unlocked.
Great post.
thank you sir again for another excellent video
Thanks for teaching me something yet again! Thanks Chris!
This was an excellent video! Along similar lines, could you do a video on mesh networking, particularly with the RPi or similar SBC? Thank you!
Ah, you are a mind reader Bill. I included mesh networks here because I have plans for a video in which I build a mesh network using several Raspberry Pis. :)
Thank you for the Sunday Lesson as always Chris! Sorry to sound facetious but doesn't Edge Computing sound rather like what we had before the cloud? :P
See my other answers here -- edge devices are still connected to the cloud.
@@ExplainingComputers I know! was being 'tounge-in-cheek' :)
This really seems to me like old 3 tier client server setups were the clients would communicate with powerful second tier server and then push back up to the main server when required. I find amazing how how we do seem to be going backward ie originally central large main frames with dul terminals then we went to powerful PCs and the 3 tiers client server then back to centralised cloud and now back to client server setups. I guess it all helps hardware providers sell more hardware in the and like retailers change the fashion to sell more clothes and how fashion goes in circles
so i'm 3 minutes in, but i think i already have a grasp of it.
Edge computing in things like Facial recognition software helps in.. well.. speed and efficiency, since Edge computing is as the name suggests, as close to the source as possible.
Very interesting stuff indeed.
Exactly right.
@@ExplainingComputers just finished the video. There was quite a bit more than that but good video either way!
Thanks for the detailed explanation
An interesting video thanks Chris. Are you planning on showing some actual edge computing demonstrations on some SBCs? I'd like to see some more.
Oh yes, I will be returning to both edge computing and mesh networks with some SBC videos. This video kind of lays down some foundations.
Very well done video, informative.
Say it: I am not a robot. Say after me: I am not a robot :D
nice video again. You made alot clear again