Instead of one post per month, I’m doing one post for January, February and March of 2010 because it’s the same story. ResNet hasn’t gotten any better, and if the outage numbers are any indication, managed to get worse.
January was notable for having 224 separate small outages, a new record, though March did come close with 214 separate outages. February clocked in at 107 outages, though it was likely lower due to the two-week Olympic break where many residents were not around (the two weeks of the break recorded 23 outages, while the other two had the remaining 84).
It appears that the ISP still has not configured their equipment properly, which in the case of a slow connection between Residence and SFU has been ongoing for over a year (the November post has more information on this issue).
Additionally, whenever the interconnect between SFU and Residence goes down, so does connectivity to SFU completely, which indicates that their router is not routing. This is in contrast with the way it should work, where a failed link is routed around, in this case going out over the ISP’s internet link and then back to SFU.
Why both of these issues continue to be issues, I can only guess at, but I suspect it is based on the ISP/Residence not caring enough to properly fix it.
Also noted were 13 outages on Jan. 4th and 25th, March 23rd and 14 outages on Feb. 5th. March 23rd also saw a few minute outage and drastically higher packet loss and latency into the evening. The reason given in was that someone had attacked the network internally, but there was no evidence provided, and given the veracity of previous statements, I would not put much weight behind it.
From my own logs and investigation, it looks more like someone, whether intentionally or not, was able to cause issues with the gateway due to poor network design/configuration.
Despite it being a preventable outage, no refund has been forthcoming, which was entirely expected due to Residence’s continued animosity towards taking steps to improve ResNet or treat their customers properly.
There was also an outage of a significant portion of an hour on the 15th of March and on the 29th of January. There was also notable instability, packet loss and latency on the evening of March 1st and 29th.
I realized when I was compiling these statistics that issues generally crop up in the evening, which is when the network is under the most load. In fact, upon more closely examining my logs, there is a definite trend in latency and packet loss throughout the day that peaks around midnight and falls off over the early morning hours, levels out in the late morning and afternoon and starts increasing around 5PM.
Which leads me to a conclusion I have previously reached, that the network capacity is too low and that the ISP is quite oversubscribed (though this is just further evidence, I reached that conclusion upon learning the 90:1 contention ratio).
Uptime was 98.1% in January, 98.9% (so close to 99%) in February and a more dismal 97.6% in March.
Downstream speed averages continued the trend, being at 10.3Mb/s in January, 11.0Mb/s in February and 9.8Mb/s in March, the first time the average speed fell below 10Mb/s, taking it out of the high-speed internet category (I consider 10Mb/s about minimum for something to be high speed, any lower than that and it’s not terribly usable for things requiring speed).
Average upstream bumped around .75Mb/s, for January, with February and March coming in at .89Mb/s and .71Mb/s respectively.
Average roundtrip time with the gateway was 38ms in January, 29ms in February, and a stunning bad 43ms in March. I say stunningly bad because it finally eclipsed the average roundtrip time on my Shaw connection, which goes to a much further distance, both in terms of the network and geographically.
Whereas Shaw’s gateway is downtown (roughly 15km away), ResNet’s gateway is on Residence, or a few hundred meters away and connected by low latency Ethernet and fiber, versus Shaw’s higher latency hybrid fiber and coax network.
As is a recurring theme, this just shows… You know how this goes, it’s overloaded.
The best summary I can come up with is that I am paying too much for too little. The ISP has not fixed problems that have been ongoing for over a year now, and Residence seems to be more interested in saving face than actually fixing issues.
