AT&T Broadband → PINGS - TRACERTS - MODEM DIAGS -Time 4 next level?

CISCO or not CISCO it's probably the way the routers are configured, and nothing to do with the brand

I guess you buy into the CISCO hype, but there are many other high quality vendors out there.

When I call Tier-1 or Tier-2, they can't even see my modem/system?!? So, they can't even check my caps or what config file my modem sucked down?!? Can people on these routers upcap their modems, while we are in this stealth mode?!? I'm not gonna try it, because I don't have another broadband provider to go to.

said by r0bErT4u:

---

I'll reclassify the routers from POS to NOT ATTBI FRIENDLY!
Probably configured correctly, but ... Cisco Diags don't work on the Cudas ... DUH!!! That's why Tier-1 & Tier-2, etc. can't see our modems/system ... DUH!!! Why can't these Cudas sustain the speeds closer to the caps?!?
»www.basystems.com/soluti ··· ex.shtml

---

This surf'n for free 'till fixed is okay, but I'd rather be fly'n near the caps & have Tier-1 & Tier-2 tools for Cuda routers.

Ok, here are some possibilities for you as to why your throughput has all of a sudden dropped this way.

1. Crappy router configuration. If ATTBI installed Barracuda routers, then almost certainly they are configured to run OSPF (Interior BGP would be overkill for this sort of thing.)

OSPF is designed so that after a topology change takes place, the change is flooded to all neighbor routers in the OSPF topology database. Once that particular packet flood has taken place, the routers in the database just exchange Hello protocol packets at a set interval so that they all know who is still up and who is down. It isn't until either a set length of time has passed (typically 1-2 hours), or until a router stops receiving Hello protocol packets from its neighbor(s) that a topology table is flooded out to the OSPF network.

Depending on how ATTBI has structured its local areas and the connection for each of those areas to the OSPF backbone, these updates can overwhelm a local network - especially when a link within an area is "flapping". From what I see in your traceroute however, this is not the case. The pings within the ATT network are all less than 20ms, which to me indicates speedy packet switching. It isn't until the datagrams reach the Level 3 network that the echo reply time spikes. I have also noticed this particular problem here in the San Francisco Bay Area.

2. Saturated links. This is the most likely cause. I have read that ATTBI is running a field trial of 256kbps in your part of the country (OR/WA), which means that are probably an insanely high amount of ATTBI users taking advantage of this trial to transfer files as much as they can before it stops.

It could also just happen to be that ATTBI is passing a lot of other types of traffic through its backbone. ATTBI uses the same IP block as AT&T Business Internet (ATT ITS), so I would imagine that a lot of ATTBI and ATT ITS traffic share the same network resources.

Now, why the Tier1 and Tier2 techs can't see your modem is another question altogether. This could be because they do not have access to whatever SNMP community your cable modem is configured to be a part of. Or, the Barracuda routers could be setup to block SNMP traffic to/from the particular portion of ATTBI's network you happen to be a part of.

Of course, this is all educated supposition on my part. Anyone else with other ideas, feel free to chime in!

g0nepostal

g0nepostal, thanks for the info. I'll pass this information to the next truck roll tech.

Here's an update. Today, a tech call me directly from the tap & node. His modem/system couldn't be seen by tech support either, imagine that. He did the OOL Speed Test and was getting a transfer rate of about 98KB/sec = 802816 bps, or 802kbps:(

We're within 90-days. Can they get a refund?!? ATTBI should call the vendor for training & support.