Great Content! Loved it! Its always a blast when learning new features of CDN, I've been accustomed with Akamai CDN and have been amazed time and time again since I first learnt about it, by how significant role it plays in any system design scenario, and how it should be utilized to the fullest (also with caution) to minimize load on your origin servers.
If a CDN utilizes independent edge servers, the amount of content that can be cached is limited. Isn't this limitation could lead to frequent evictions, ultimately undermining the primary purpose of the cache?
There's a name for this pattern, due to which multi-level caching is required. It's called a thundering herd problem. Based on scenarios, multiple ways we can handle this. For example, Facebook uses Memcached to cache data, but it also configures some threshold, so that not all requests, goes to DB. It can also be done using a Server director pattern.
Hey Arpit! The availability of edge servers is limited by the availability of the shield servers. Further the geographical "nearness" is impacted since multiple edge servers map to single shield server. How do they get around this?
You missed talking push vs pull type of CDN
becoming my go to channel for system design
Awesome Content as always! ❤️
Great video. Thank you for sharing the link to the engineering blog as well. Your video made it easier to go through the blog.
Where did he share the link in description? I don't find it.
Great Content! Loved it! Its always a blast when learning new features of CDN, I've been accustomed with Akamai CDN and have been amazed time and time again since I first learnt about it, by how significant role it plays in any system design scenario, and how it should be utilized to the fullest (also with caution) to minimize load on your origin servers.
Man you have such an excellent way of putting up things!
Thank you 🙌
Hi Arpit ,
Can you please the prerequisites to fallow your provided video as beginner
If a CDN utilizes independent edge servers, the amount of content that can be cached is limited. Isn't this limitation could lead to frequent evictions, ultimately undermining the primary purpose of the cache?
There's a name for this pattern, due to which multi-level caching is required. It's called a thundering herd problem.
Based on scenarios, multiple ways we can handle this. For example, Facebook uses Memcached to cache data, but it also configures some threshold, so that not all requests, goes to DB.
It can also be done using a Server director pattern.
where is the engineering blog please
Sir
we say that video is just combination of images and sound
but what about gif?
how gif are made programmatically?
I'm still wondering why your subscribers are so low, your channel deserves more.
Haha. Even I wonder the same 😅 I have covered some really interesting and useful topics.
The Internet is beyond my understanding 🙃
That's cause he doesn't share the original link of the blog in the description.
Can the edge server pull and cache contents from nearest edge server if needed?
thanks for the content :]
does the structure of edge + shield + origin form kind of binary tree? otherwise sheild servers will be overloaded for viral content
Amazing content.
As a final year student, what subjects do I need to have a good grasp on to best understand these concepts of system design.
Hey Arpit! The availability of edge servers is limited by the availability of the shield servers. Further the geographical "nearness" is impacted since multiple edge servers map to single shield server. How do they get around this?
Shield servers are localised for a set of local edge servers.
If sheild servers are down edge can always fallback to origin.
Awesome 😎
Can we leverage CDN functionality in microservice architecture ?
So long as the service is user facing
Did they use rust ?