About a week ago, speeds from a private server held with OVH in Beauharnois took a tank during primetime hours (400kbs) and are not so good even during off hours(2MB/s). This prompted a fair bit of investigation on my behalf. It seems that the routing Shaw uses with OVH has changed as of about a week and a half ago.
Data used to be routed to Chicago, where Shaw interchanged the data via Equinix. According to peering DB, Shaw has a respectable 30gbps of bandwidth at this interchange (and OVH has 100). However, since the changeover, data is now routed through TORIX, where Shaw seems to only have 10gbps of bandwidth at this interchange.
MTR of the new route.
In addition to lower bandwidth overall, packet loss now seems common and delays of *several seconds* (2.8 second ping max at the time of writing) are far more common by the time data reaches the dx3ca-te3.vc.shawcable.net ix.
Is there anything that can be done to address these issues? For instance, higher bandwidth through TORIX or routing data back through Equinix where the exchange has more leeway.
It's worth noting that all data through OVH comes to Shaw through this route -- even data that originates from their Roubaix datacenter and all servers (both business and personal) at BHS. 10gbps hardly seems like enough bandwidth for all of Shaw customers from a major outbound provider like OVH and it shows clearly with the speeds seen in late evening.