Is this old news?
macresource.com
<Greg Barry has confirmed that US Robotics X2 56K modems have trouble maintaining a
v.34 connection (33.6K or less) when connected to certain modems; his letter explains
the details, and notes that Earthlink has decided to nationally support the 56KFlex
modems, and not the X2 technology.>
Here's a copy of the letter the above statement refers to.
<X2 56K Modem Bug
From Greg Barry, November 19, 1997
Greetings,
I have an alert regarding the US Robotics X2 56K modems. After numerous
phone calls involving Earthlink and USR, it seems that there is a known
problem with the X2 series of modems keeping a v.34 connection (33.6K or
less).
The problems stems from their implementation of the V.42 protocol suite
which provides error correction and compression. The symptom is that
after a few minutes of high activity after a successful connection, the
USR X2 modem keeps the connection, but no longer exchanges data with the
other modem. The line does not drop, but the modems go out of sync and no
data is transferred.
The crux of the problem is that the error occurs only with certain
modems, and of course not between USR X2 modems. Since Earthlink uses
UUNET and PSINET for it's national POPs, I decided to try some other
numbers around the country - same thing. UUNET and PSINET must use the
same vendor for all of their modem banks.
What is causing this problem is the fact that USR has adhered tightly to
the V.42 standard, whereas most other modem vendors have not been so
narrowly focused. This wide focus allows modems from different vendors to
work together.
While speaking to a level 2 tech support engineer at Earthlink, he gave
me the lowdown on the issue and directed me to USR level 2 support to get
the fix. Another interesting wrinkle to this is the fact that the
Earthlink engineer mentioned that Earthlink had decided to nationally
support the 56KFlex modems and not the X2 technology. Besides the
technically superiority of KFlex, he said that they received thousands of
calls from X2 modem owners with this same problem and that a national
roll-out of X2 would be a nightmare.
USR has admitted to the problem (I have a call reference #) and says that
they will be addressing it in a flash upgrade "someday." Their short term
fix, besides returning the modem for another brand, is to disable V.42
with S15=128. This greatly reduces the performance and allows for errors
to corrupt the datastream.
Coupling this issue with the rumors that the X2 modems will not be ITU
upgradable, it seems that USR is in some deep trouble and as usual is not
owning up to any of it (without a great deal of prodding).
Greg Barry>
Perry |
