OVH network capacity has hit 15 terabits per second!
OVH has deployed, operated and maintained its own fiber optic network since 2006. It connects 32 points of presence (PoP) around the world. How does this infrastructure benefit our customers? Why do we need to keep developing it continuously? And what does “network capacity” mean? Read on to find out.
What exactly does “network capacity” mean?
All of our 28 data centers are connected to the internet via the OVH backbone. By this, we mean our own network, operated by our own teams. This network spans across 32 points of presence (PoPs), which are connected to each other via fiber optic cables.
Each PoP acts like a highway intersection between OVH and other providers. These might include internet service providers, cloud providers and telecom operators. When we talk about network capacity, we mean the total capacity of all of these “intersections”. And that is what has reached 15 Tbit/s.
Having our own network is a major advantage in our business. It gives us far greater control over quality, since we are as close as possible to our customers and their end users. In contrast, most of our competitors rely on a third-party “transit” operator. As a result, they have less control over quality.
Why do we need such massive capacity?
By managing our backbone, we can offer our customers high-quality bandwidth and low levels of latency anywhere and everywhere in the world. It also helps us to protect users against distributed denial-of-service (DDoS) attacks.
In fact, one reason OVH needs a lot of excess capacity is to absorb the spikes of high-intensity attacks. Because it’s not just about knowing how to handle the attacks, but about being able to withstand them. It is clear from our 2017 figures that DDoS attacks continue to increase in volume.
As a cloud services provider, we naturally have a lot more outgoing traffic than incoming traffic. And it keeps growing by around 36% per year. Nevertheless, it is still essential for us to increase our capacity for incoming traffic, if only to avoid congestion caused by DDoS attacks. This is how we are able to manage and filter attacks using our anti-DDoS VAC system without legitimate traffic suffering any saturation issues.
The excess capacity also increases availability. If, for example, we have an incident in one of our PoPs, or one of our partners (e.g. an operator or internet service provider) experiences a fiber optic cable outage, it doesn’t matter. We can still avoid any congestion, and maintain our service quality in terms of low latency. And we can do this because we are closely connected to our partners via several PoPs. If we are connected to the same operator via two PoPs, that means we can’t exceed 50% usage for each PoP. If we are connected via three PoPs, that means 33% each, and so on. In the event of an incident or a failure, the main difficulty is anticipating where the traffic will be redirected (in other words, to which PoP).
Spotlight on upgrades to the Asia-Pacific network
Our recent network upgrades in the Asia-Pacific region really demonstrate how OVH teams can intervene at all levels of the backbone, to improve availability and help protect customers.
New core routers have been added to the data centers in Sydney and Singapore. Each router supports up to 4.8 Tbit/s of routing capacity.
We have also installed four new backbone routers (two in Singapore and two in Sydney), which are all connected via 100 Gbit/s links. With these additions, OVH can connect with more providers, and as a result we can expand our capacity beyond our own network.
OVH has also invested in 100 Gbit/s links to connect with other PoPs. Increasing our capacity to handle DDoS attacks has strengthened our network’s security.
Finally, we have been able to gain precious milliseconds by adding a new low-latency link between two OVH data centers in the Asia-Pacific region (88ms latency instead of 144ms), and modifying the routing via Singapore.