Under attack?
Upstreams & peering

Ten ways to reach a destination, and logic that picks between them

We are a Tier 2 network and say so plainly. A Tier 1 carrier sells you its single view of the internet; we buy from ten of them and peer at twenty exchanges, which means there is a choice to make for every destination - and something making it continuously rather than accepting whichever AS path is shortest.

Upstreams & peering

A premium transit mix, chosen for latency rather than for price

Cheap transit is easy to buy and you can hear it on a voice call. Our blend is built from Tier 1 carriers, internet exchanges and private peering, and the routing between them is decided by our own path-selection logic rather than by whichever AS path happens to be shortest.

5 TbpsTier 1 transit capacity
15 TbpsInternet exchange and private peering capacity
50 TbpsBackbone capacity between points of presence

We are a Tier 2 network, and that is deliberate. A Tier 1 carrier reaches the whole internet on its own peering and sells you that single view of it - one set of paths, one set of disputes, one set of congested ports on a bad day. We buy transit from ten of them and peer at twenty exchanges, which means that for any destination there are several ways to get there and we can choose between them. The routing logic below is what makes that choice, continuously.

Low latency

The shortest path that is actually fastest

BGP picks the fewest autonomous systems, which is not the same thing as the quickest route and frequently is not even close. We measure the paths we hold and prefer the one that is performing, not the one that looks tidiest in the table.

Zero packet loss

Congestion is a routing decision

Loss on a transit path is almost never a fault - it is a peering dispute or a saturated port somebody else is not in a hurry to upgrade. Holding ten upstreams means we can route around it the same day rather than opening a ticket and waiting.

Our own selection

Path logic we wrote ourselves

Continuous measurement of latency, loss and jitter per destination network across every available path, feeding a preference that updates automatically. The same instinct as the filter: measure it, then decide - never assume the default is right.

Tier 1 and transit upstreams

NetworkASRole
GTTAS3257Tier 1 transit
NTTAS2914Tier 1 transit
CogentAS174Tier 1 transit
Tata CommunicationsAS6453Tier 1 transit
OrangeAS5511Tier 1 transit
Hurricane ElectricAS6939Tier 1 transit
SparkleAS6762Tier 1 transit
ArelionAS1299Tier 1 transit
LumenAS3356Tier 1 transit
RETNAS9002Tier 1 transit

Internet exchanges & private peering

AMS-IXDE-CIXLINXGNM-IXBIX.BGFRYS-IXSPEED-IXGigaNetInterLANGlobal-IXNINE-IXPiter-IXERA-IXLSIXpeering.czSIX.SKGR-IXEquinix IXInterIXTurkIX

Across them we hold thousands of peering sessions - the exchanges are where the long tail of the internet is reached directly.

Private interconnection

The heaviest destinations do not ride transit at all. We run hundreds of direct PNI ports into the content, cloud and hosting networks that carry most of the traffic - a private cable to each, at each point of presence where they meet us:

Google Meta AWS Microsoft Apple Netflix Cloudflare Akamai Fastly CDN77 ByteDance OVHcloud Hetzner Leaseweb DigitalOcean

A PNI is a port nobody else's congestion can touch: no shared exchange fabric, no transit path, one cable between two routers. It is why traffic to these networks stays flat on the evening the rest of the internet is busy.