A month of Astiga's new synchronisation system
Koen van Hove is a software and research engineer at NLnet Labs and a guest PhD candidate at the University of Twente on the topic of network security and responsible disclosure. His most notable contributions are to the DDoS Clearing House and RPKI Relying Party Resiliency Platform. He is also a researcher level 2 at the Dutch Institute for Vulnerability Disclosure (DIVD), active within the Internet Engineering Task Force (IETF) and RIPE community. He writes about technology, policy, and other random things that catch his interest.
This is a follow-up to the previous blog post, which you can find here.
Astiga's new synchronisation system has run for about a month now. Which seems like a nice time to evaluate it. Some stats:
About 50 TiB has been synced in a month, which is a bit under 2TiB per day. That also has to do with the start, as I expect to break the 60TiB barrier for March 2020. Compared with the old system, that is about double the data it used to process in a month. Apart from the first initial hickups (nothing too noteworthy), everything now works fully as it should.
Something I have noticed is that the original configuration was a bit on the careful side - there is still room for an increase in the amount of slots (and thus an increase in speed). This will be gradually done in the future.
Nothing really special happened, everything just works as it should. Things that just work as they should don't tend to make for interesting blog posts, so my apologies for that - next time I will make sure something crashes in spectacular fashion :-)