I have decided to switch to Sonic from Megapath (was a happy customer of speakeasy for years before they were acquired by megapath, and after the most recent week-long outage, it's time to find a new provider).
I am in San Francisco. I am a software engineer, and I spend most of my time connected via ssh to machines hosted in google's cloud. Because I spend most of my typing, I care more about latency and packet loss than bandwidth. It's very difficult to write code with high or variable latency (and packet loss makes it basically impossible). Switching to a new ISP is risky because it's very hard to predict ahead of time what the performance will be like. Can you tell me a little bit about how you guys operate in general that might make me more confident about the switch, or alternatively, can you tell me what kind of performance I can expect getting from you to google's network? (google has tons of peering, but my current traceroute reaches google at eqixsj-google-gige.google.com (206.223.116.21))
Here are some random questions. Feel free to answer any of these or perhaps just say a few things about how you operate that might inspire some confidence ...
Do you guys run your own backbone?
Do you have multiple peering relationships?
Do you generally send traffic over free peering points?
Do you guys stay ahead of the curve in terms of provisioning, or is packet loss (or high latency /queueing) something I should expect to experience at peak hours?
Do you operate a NOC?
If there is a network problem can customers speak directly with the NOC?
Do you monitor your network for latency and packet loss?
Do you monitor it for traffic congestion?
Do you monitor your peering points for congestion?
Do you actively re-route traffic when upstream providers exhibit poor performance, or do you rely entirely on BGP?
I am in San Francisco. I am a software engineer, and I spend most of my time connected via ssh to machines hosted in google's cloud. Because I spend most of my typing, I care more about latency and packet loss than bandwidth. It's very difficult to write code with high or variable latency (and packet loss makes it basically impossible). Switching to a new ISP is risky because it's very hard to predict ahead of time what the performance will be like. Can you tell me a little bit about how you guys operate in general that might make me more confident about the switch, or alternatively, can you tell me what kind of performance I can expect getting from you to google's network? (google has tons of peering, but my current traceroute reaches google at eqixsj-google-gige.google.com (206.223.116.21))
Here are some random questions. Feel free to answer any of these or perhaps just say a few things about how you operate that might inspire some confidence ...
Do you guys run your own backbone?
Do you have multiple peering relationships?
Do you generally send traffic over free peering points?
Do you guys stay ahead of the curve in terms of provisioning, or is packet loss (or high latency /queueing) something I should expect to experience at peak hours?
Do you operate a NOC?
If there is a network problem can customers speak directly with the NOC?
Do you monitor your network for latency and packet loss?
Do you monitor it for traffic congestion?
Do you monitor your peering points for congestion?
Do you actively re-route traffic when upstream providers exhibit poor performance, or do you rely entirely on BGP?