Some users on the virgin media forums say virgin purposely used the puma chipset since its latency issues help to take some stress off the virgin network itself. Any take on this theory?
It's usually the sort of illogical post made by people who have no grasp of how time, supply agreements or Liberty Global's choices dictate the hardware VM use.
The bug in question relates to Texas Instruments time, Intel purchased them and re-branded the chipset, it was still unknown when Intel released it's own branded SoC's, same when Arris integrated them and again when Liberty Global awarded them the supply contract across all? of it's markets. So TI cocked up, Intel cocked up, Arris cocked up, LG cocked up and VM has near zero input on hardware it gets given. Does that sound like they chose anything, let alone did so to reduce network load when they've over provisioned and increased speeds multiple times for affected customers over the years? I mean even if the flaw came to light early, the supply contract would have been agreed well before the products first launched and VM have no way of dodging that bullet. I also doubt intel had any awareness, if they did, they would have quietly notified partners and released a firmware fix ahead of disclosure, they took months and months along with multiple attempts to fix the issue and it's still only mitigated.
The whole situation is a mess, but VM is stuck with the Hub4 and potentially further generations (though i'd be amazed if Intel hasn't sorted it's SoC's by then). Even if VM identify the issue during internal testing (and having been involved in closed internal trials and UAT with VM, I would love to now what they're playing at not to do basic throughput tests using the two most common protocols in use), they can only feedback the issue and hope it's raised with Arris who would then confirm it's a SoC hardware issue and pass back to intel.