31–40 of 43 posts
I wonder what technology they use under the hood. Ceph? GlusterFS? Custom software?
they chose to make their own stack afew years ago instead of going with the already available and business friendly openstack, going by the same reasoning and the long amount of time they took to reach here, it highly likely to be a custom thing.
Slowly, very slowly, DO is creeping towards having enough features to peel users off of AWS and GCP. Block storage, load balancers, and now blob storage...if you don't need managed queues or databases, you can probably save some serious coin by running services on DO. S3 sets a high bar in terms of durability and availability, so it will be interesting to see how well DO can compete, and it will take a long time to g…
Earlier quoted context omitted.
they chose to make their own stack afew years ago instead of going with the already available and business friendly openstack, going by the same reasoning and the long amount of time they took to reach here, it highly likely to be a custom thing.
If OpenStack is "business-friendly" then it's very particular about who its friends are.
Slowly, very slowly, DO is creeping towards having enough features to peel users off of AWS and GCP. Block storage, load balancers, and now blob storage...if you don't need managed queues or databases, you can probably save some serious coin by running services on DO. S3 sets a high bar in terms of durability and availability, so it will be interesting to see how well DO can compete, and it will take a long time to g…
I wouldn't be surprised to see DO either roll out DaaS next or partner with someone (Compose.io?) to do it. That's one of two things (the other being VPC, which I know they're working on per HN posts elsewhere) that make AWS/GCP tempting over DO for me. That said, I kind of hope it's not Compose.io - they seem ridiculously overpriced for PostgreSQL compared to AWS's and GCP's offerings.
I use DO personally for small things and like it a lot, but VPC is a hard requirement for company work.
Earlier quoted context omitted.
Good to know. Am in the throes of choosing between both for a simple project
VPC is in the works at the moment, still a bit too early for a release date, but the team has been on it for about the past 3 quarters.
Earlier quoted context omitted.
I wouldn't be surprised to see DO either roll out DaaS next or partner with someone (Compose.io?) to do it. That's one of two things (the other being VPC, which I know they're working on per HN posts elsewhere) that make AWS/GCP tempting over DO for me. That said, I kind of hope it's not Compose.io - they seem ridiculously overpriced for PostgreSQL compared to AWS's and GCP's offerings.
VPC on DO is the #1 feature request that stops my team from considering it. I use DO personally for small things and like it a lot, but VPC is a hard requirement for company work.
Earlier quoted context omitted.
VPC is in the works at the moment, still a bit too early for a release date, but the team has been on it for about the past 3 quarters.
Seed-stage startup here that would love to start using more of our (time limited) DO credit where our cluster requires running in a VPC. Any way to sign up for early access to VPC when it does roll out?
Earlier quoted context omitted.
Is there a stand for object storage?
openstack swift supposedly but this is a slightly incompatible S3 implementation that nothing else talks. The filesystem works pretty well though ;)
Earlier quoted context omitted.
they chose to make their own stack afew years ago instead of going with the already available and business friendly openstack, going by the same reasoning and the long amount of time they took to reach here, it highly likely to be a custom thing.
If OpenStack is "business-friendly" then it's very particular about who its friends are.
1 - First was that it didn't really work. You could stand it up but DHCP licenses would fail, your VMs would go down, it just wasn't quite stable.
2 - Naively we thought it was a bit too complicated. Sometimes early on being naive is great and we certainly made a more simple system ourselves, however as time went on we realized, that our backend was beginning to resemble OpenStack in complexity, but we would still have more flexibility in hiding that complexity from our end users.
3 - OpenStack is designed for organizations but not necessarily to be multi-tenant. Taking any software and making it multi-tenant for different customers is a large effort, so coupling that with OpenStack not being mature, just seemed like a lot of effort to put into it.
4 - It wasn't really Open Source the way we were used to it. We were used to organic open source efforts by single developers or teams of developers that naturally grew and developed over time. OpenStack just looked like it was very much "corporate" sponsored open source, which wasn't something we felt comfortable putting our faith into.