Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[GigabitEthernet]

The Huawei HG633 is listed by this very website as being SIN 498 compliant however, Huawei/TalkTalk have literally just started rolling out a firmware that supports ReTX/G.INP.

Retransmission for Fibre customers (customers who’s connections require it will now be able to take advantage of the ReTX technology)

How then, could this router have passed?

[gt94sss2]

The reason the HG633 passed MCT is that it supported g.inp before the new firmware that you mention above and Openreach were able to successfully test for it during their testing.

http://www.talktalkmembers.com/t5/Superpowered-Fibre-Broadband/TalkTalk-Super-Router-Huawei-HG635/td-p/1636235 has TT confirming it back in Mach 2015.

I suspect all this new firmware does is tweak the implementation somewhat.

[GigabitEthernet]

That's confirming the HG635 as supporting it

The HG633 did not support it until today I do not think. I previously had G.INP on my line in the upstream and downstream directions. Upon putting on the HG633, the next morning my interleaving depth was put up over 1000 on the downstream and something like 50 on the upstream side, by DLM.

Upon being put on the v1.20t firmware, DLM reduced the interleaving depth on the downstream side down to 5 and the upstream down to 1, to me at least indicating that G.INP had been re-enabled.

[kitz]

Looks like TT may have been using their own version of f/w, which may not necessarily be the same one as tested.

The Huawei f/w used during MCT was G3_01.32.10.
Note they say in that TT thread firmware v1.20t.  Previous versions were also in the format v1.18t etc.. indicating they use a custom/bespoke version.

[ejs]

But aren't all devices supplied by ISPs supposed to have passed the testing? The list of approved devices we've seen has never included any device custom built for an ISP, it's a list of only non-ISP-branded devices.

The only idea I had about how it could have passed the testing but not worked properly on the actual network, was that the line profile used for the test isn't the same as the profile actually used by the DLM on the network.

[kitz]

Yes, Yep... and possibly.

[ejs]

It's occurred to me the the firmware versions given in the list of approved devices is the usually the modem firmware, not the device firmware as a whole, which may account for the differences in version schemes. It also occurred to me that a device doesn't necessarily need to be re-tested if the modem part of the whole device firmware is unchanged, so the TalkTalk firmware might not need to be tested separately, if they don't change the modem firmware and driver.