A next generation network designed to perform better than the Internet
PacketExchange provides a service portfolio running over our Layer 2 Ethernet over MPLS network that performs incrementally better than the public Internet.
Everything we do will be guided by the goal of making sure that every packet sent is routed in the best and most efficient way possible and over the most reliable and secure infrastructure available. This way, we believe that the online strategies of every organization, big or small, can be improved upon. Both of these headline roadmaps will further enhance the capability to deliver traffic between originator and respondent with guarantees of performance, reliability and security.
|
Typical Internet |
PacketExchange Network |
| Downtime - Frequency |
Probably 2 unscheduled outages per site per year. If software/DDoS, there can be 2 or 3 failures in succession. |
Less than 1 per site per year |
Downtime - Impact
|
If static routed, customer will lose all external connectivity.
If BGP, connectivity via that provider is out for the duration. |
Where diversity is deployed and purchased by customer, 95%+ of failures will be partial. |
Downtime - Duration
|
Typically 3-6 hours. |
We meet our 4 hour restoration time 95%+ of the time (and most failures will be partial). |
Re-convergence
|
3-5 minutes |
50ms restoration available everywhere |
Latency
|
Unpredictable. Can be equivalent to a private network, but provider can re-route traffic at any time which can result in increases of up to 50ms. |
Small progressive performance improvements as meshing increases. Typical performance metrics are:
London - New York = 70ms RTT
New York - Los Angeles = 74ms RTT
Ashburn - Seattle = 79ms RTT
|
Jitter
|
Not consistent: Ranging from 0.3ms to 50ms. |
<0.1ms Jitter |
Packet Loss
|
This is dependent on destination and can be as bad as 5-6%. If a bad route is selected for a destination, this is very hard to get fixed. |
<0.1% |
