#60: Break Into Google Spanner Architecture (5 Minutes)
I didn't know they were using Paxos!
Google has always been a pioneer in high scalability challenges; excellent study case, Neo👌
thank you very much.
I think that’s the first time I’ve read how Google Ads was built. 🙌
nice, thanks.
Hi, great article ! Could you please tell what is the application you used for creating diagrams?
hey Georgiy, I used Eraser to create block diagrams:
https://app.eraser.io/auth/sign-up?ref=neo
So, does the follower compare the last stored record with the leader's data?
only the timestamps.
Wouldn’t be inefficent for a follower to reach a leader in a different zore in terms of performance?
What should be the approach here?
good question, Ata.
I think it's about the tradeoffs. Some clients could tolerate eventual consistency, thus achieve relatively low latency reads.
Also the leader could be deployed in zone where most users are located for better performance.
This is really insightful. Great read. I have thoroughly enjoyed reading it
thank you so much, happy to read your feedback.
How exactly the cloud spanner db handles scalability while also being a strong RDBMS?
undefined subscriptions will be displayed on your profile (edit)
Skip for now
For your security, we need to re-authenticate you.
Click the link we sent to , or click here to sign in.
I didn't know they were using Paxos!
Google has always been a pioneer in high scalability challenges; excellent study case, Neo👌
thank you very much.
I think that’s the first time I’ve read how Google Ads was built. 🙌
nice, thanks.
Hi, great article ! Could you please tell what is the application you used for creating diagrams?
hey Georgiy, I used Eraser to create block diagrams:
https://app.eraser.io/auth/sign-up?ref=neo
So, does the follower compare the last stored record with the leader's data?
only the timestamps.
Wouldn’t be inefficent for a follower to reach a leader in a different zore in terms of performance?
What should be the approach here?
good question, Ata.
I think it's about the tradeoffs. Some clients could tolerate eventual consistency, thus achieve relatively low latency reads.
Also the leader could be deployed in zone where most users are located for better performance.
This is really insightful. Great read. I have thoroughly enjoyed reading it
thank you so much, happy to read your feedback.
How exactly the cloud spanner db handles scalability while also being a strong RDBMS?