Looks like a bandwidth issue to me, since even doing the download manually, my connection stalled 5 times before I could complete a successful download. And the download speeds were atrocious, many times in bytes/second rather than even kb/second - and my connection speeds to the Internet are in multiple 100mb connections.
Have you considered mirror sites or adding bandwidth?
Normally our bandwidth is sufficient. We have considered mirror sites also, and we have plans to move our hosting into a local Equinix facility where we will have similar bandwidth to yours and other benefits. Unfortunately we are not quite up to that level of revenue yet.
We currently have two T1s through two networks (Savvis & Sprint). More than 90% of the time more than 80% of our bandwidth is avaialable. There are occasional short-lived peaks where this is not the case, but those are rare.
Rulebase compilation is metered so that each file is generated in about the same amount of time it takes to download the file through a single T1. Generally this pacing leaves our bandwidth mostly open most of the time.
However, it appears that something odd has been going on recently with the Sprint side of the network - I suspect that what you've observed is related to some flapping going on under some heavy load conditions and that this has led to a number of dropped packets. I am investigating this further.
An event such as this would reduce our bandwidth by more than half and many packets would be lost.
_M
This E-Mail came from the Message Sniffer mailing list. For information and (un)subscription instructions go to http://www.sortmonster.com/MessageSniffer/Help/Help.html
