In May 2013, we have made a major decision for the future of alwaysdata. After taking nearly nine months to research, compare, calculate, purchase, install, develop, debug, and document, it’s now time to announce it.
Our own infrastructure
Since its inception in 2007, alwaysdata has relied on third-party providers (mainly OVH, Online, Hetzner and Amazon) for its infrastructure (hardware, network). This has allowed us to concentrate on the heart of our architecture: the software. alwaysdata would not exist without these providers; we owe them a lot.
Not having our own infrastructure has many advantages, but it also has drawbacks:
- Part of our reliability depends on our providers
- We are limited to equipment and features offered by our providers
- We are subject to changes in their service policy
As our infrastructure grows, these drawbacks become more important. After a long discussion, we have decided to take the plunge and become independent.
We have two objectives:
- Build a reliable and durable architecture
- Be completely independent
These objectives result in many choices that must be made. Here are the key ones:
- We are hosted in a neutral data center (operated by Equinix), without intermediary
- We joined the RIPE (we are “LIR“) to ensure our independence
- All our equipment is doubled in power
- Our internal network is fully redundant: each server has two network connections (of different brands), which are themselves connected to two different routers (of different brands)
- We have three different network operators (transit)
Some additional technical details:
- Data center: our equipment is located in the Equinix PA3 data center in Paris
- Network operators; IELO, Cogent and Interoute
- We are on the Equinix IX Paris peering point
- Our AS number is 60362
In practice, what changes?
For our customers, nothing changes directly: our service and our price remain the same.
Our current servers, located with our providers, will be moved to our infrastructure progressively over time — the total migration is likely to take up to two years. We will announce each migration beforehand. Our dedicated customers will be contacted personally and we will organize the whole server migration.
In a few weeks, new accounts will be hosted to our own infrastructure. For new dedicated servers, this is already the case.
The real changes will be visible indirectly; some examples:
- We will be able to offer new services. Did you know that Equinix operates in 14 countries?
- The equipment we use is more suitable for our purposes. For example, the drives and power supplies of our servers can be hot swapped
- The range of equipment that we can use is broader — for example, SSD or drives with very high capacity
- Our network, fully redundant and simplified, should be more reliable
We spend much of 2013 in the construction of our infrastructure, which has somewhat slowed the pace of new developments. We are planning several things for 2014 — our team has also welcomed a new developer, Martin, a few weeks ago. Our goal is to provide more reliability, more features, and to remain available to quickly answer all questions and concerns from customers.