Serverless Architecture, Microservices and AWS Prime Monolith
Вставка
- Опубліковано 15 тра 2023
- The move from a distributed microservices architecture to a monolith application helped Amazon Prime team achieve higher scale, resilience, and reduce costs.
In this video I will look back into the basics of microservices and serverless architecture, how they have been used by different companies and if its a good idea for you to move to a monolith based on AWS Prime team use case.
#aws #serverless #monoliths
References:
1. www.primevideotech.com/video-...
2. m.signalvnoise.com/the-majest...
3. Dropbox: www.geekwire.com/2018/dropbox...
4. Netflix Outage: www.reuters.com/article/us-ne...
System Design Interview Basics Playlist:
► • System Design Intervie...
AWS Certification:
►AWS Certified Cloud Practioner: • How to Pass AWS Certif...
►AWS Certified Solution Architect Associate: • How to Pass AWS Certif...
►AWS Certified Solution Architect Professional: • How to Pass AWS Certif...
great explanation, thanks for sharing!
monolith servers can be used to create great impact, but microservice arch also has various other use cases.
Yes, exactly
The saving had nothing to do with Microservices vs Monolith, they have made other optimisations and used different technologies e.g. VPS vs Lambda, VPS's typically perform better. If they replicated exactly what they have in a monolith this would be a different story. I feel the initial design was just bad design and there is still a lot to optimse, like the processing of many frames as it is being streamed, just offload it to the client as the client knows if frames has been skipped. This has been covered many times.
I agree with the comment that the savings Amazon Prime Video achieved by moving from microservices to monolith were not solely due to the architectural change. There were other factors at play, such as the use of different technologies and the optimization of the codebase.
VPSs (virtual private servers) typically perform better than Lambda functions, which are serverless compute functions. This is because VPSs are dedicated resources that can be provisioned with more memory and CPU power. As a result, they are better suited for applications that require a lot of processing power.
afaik, Amazon Prime Video team also made other optimizations to the codebase, such as offloading the processing of frames to the client. This means that the client is responsible for decoding the frames and displaying them on the screen. This can reduce the load on the server and improve performance.
Overall, I believe that the Amazon Prime Video team made a number of changes that contributed to the savings they achieved. The move from microservices to monolith was one of these changes, but it was not the only one
@@ByteMonk it wasn’t from the move from Microservices to Monolithic though, for starters the. Are not monolithic still, even if they say they are. VPS’s are not dedicated, they are shared resources, though you get more power. The savings are from optimising bad code and decisions from a rushed build of an application. The frames are still processed in the backend.
Your audio is clipping really badly in the intro. Turn down the gain on your audio next time
I misconfigured audio settings while changing audio devices :( , lesson learned, thank you 🙏
1. Step functions is not the best product out there for workflows. A different workflow solution which implements the SAGA design pattern (or custom implementation) would be better.
2. Why did they use S3 instead of EFS again? Did they go all-aws with their implementation? Also, were they moving the files into and out of S3 using the public internet, instead of a vpc endpoint? Sounds really bad design on many aspects.
Step Functions is a powerful workflow orchestrator that can be used to implement a wide variety of workflows. It is scalable, reliable, and cost-effective. However, it is not the only workflow orchestrator available, and there may be other solutions that are better suited for a particular workflow. For example, if the workflow needs to implement the Saga design pattern, then a different workflow orchestrator may be a better choice.
S3 and EFS are both object storage services offered by AWS. S3 is a more general-purpose storage service, while EFS is designed for file sharing. There are a few reasons why someone might choose to use S3 instead of EFS for file storage. For example, S3 is more scalable and can be used to store files of any size. Additionally, S3 is more widely available, as it can be accessed from anywhere in the world.
Whether or not using S3 instead of EFS is a good design decision depends on the specific requirements of the workflow. If the workflow needs to store large files or needs to be accessible from anywhere in the world, then S3 may be a better choice. However, if the workflow needs to share files between multiple applications or needs to provide high performance, then EFS may be a better choice.
Regarding, Were they moving the files into and out of S3 using the public internet?
This is not a good practice, as it can expose the files to security risks. A better practice would be to use a VPC endpoint to connect to S3. A VPC endpoint is a private network address that allows you to access AWS services without exposing your traffic to the public internet.