Around January 1, we will be replacing our 10 GIgE circuit to Savvis in Chicago with a 10 GigE circuit to NTT in each Palo Alto, Chicago, and New York. This will increase our overall network capacity and scope, improving service out of our new POPs. This decision is also being made to improve overall network performance globally. There should be no major network outages/issues relating to this change.
We have been using Savvis ever since we opened our dedicated server and colocation operations back in 2004. THis is not a decision we are taking lightly, but over time, we have simply seen a lack of overall network improvement and growth. Savvis target seems to be largely on the data center side, where they compete with us, and in gaining financial and enterprise transport business, which does not help us as far as network reach. In addition, Savvis' network outside of the US has been relatively weak and does not seem to be getting any stronger.
On the other hand, we have been watching the continued growth of NTT while they increase their overall network reach, as is demonstrated by the Knodes index: http://www.fixedorbit.com/metrics.htm. Being a Japanese based copmpany, they have a significant Asian network, and with the Verio roots there is also significant US and European infrastructure. According to representatives at NTT, they also plan to make a push into South America, helping us to connect with those emerging markets.
This change will result in various route changes and I am certain some of you may see slightly increased latency, though based on our own measurements the improvements we see should far outweight the negatives.