atcomsystems.ca/forum Forums VOIP Phones & IP PBX Networking Point-to-Point dropping

Hello folks, I'm in a bit of a pickle and looking for some sage advice. Here's the deal:

I've got an office in Manhattan that connects to a warehouse site in New Jersey via a point-to-point T1. This link supports both voice and data. It uses RAD WebRanger routers on each end connected to Adtran TSU120s. The problem I'm experiencing is frequent data drops on the circuit, particularly when load increases to more than a few (read: more than 3) users. I also see drops when we send print jobs to the network printer at that site. By drops I mean ping timeouts when pinging from my HQ site to the router in Manhattan. I am not seeing ping timeouts to the connecting router in New Jersey.

By way of background, we use Citrix Presentation Server, so printing traffic from stations at that site would still go over the WAN connection.

I can logon to the remote routers from where I'm at, but I cannot access the Adtran boxes. On the routers I do not see any errors. What I do see is that the packet counters reset whenever the connection drops. I'm not too familiar with the WebRangers, but on Cisco routers this would generally happen with a power reset.

It's interesting to note that when I do see a drop in packets, the duration of the drops is very consistent. That is, I see 5 packet drops when pinging remotely each time there is a loss of connectivity.

Also of interest is the fact that these drops will also occur between a PC in Manhattan and the router out there. That is, I can see pings dropping remotely to the router and I also see pings locally drop at the same time. Also, I have been on the phone (remembering this is a voice and data circuit) and have not noticed any loss in voice continuity despite an approximately 25 second drop in data.

Given all this and the fact that I've already replace a small switch out there that was overheating, I figured perhaps a power issue. I sent a UPS to the site to see if that would help. No luck, though we'll connect to that box to see if there are any errors logged. It doesn't seem to make sense that power would be the only issue since simply increasing line load can reproduce the issue on the data side.

I have had the provider do head to head testing to no avial. They see the circuit as clean. I'm not seeing an errors on the router, but as I mentioned, the counters reset after each outage.

One of my coworkers is making a site visit tomorrow and will physically inspect the environment. We're not at all familiar with the Adtran units, but will bring a manual and serial cable to take a look.

We're pretty stumped, though the though of a timing issue does come to mind. Anyone have an idea on what my tech might look at while up on-site?

I'm getting closer (I think) to the root of the problem here. My tech has spoken with Adtran folks and some techs that support our phone systems in NJ/NY.

The Adtran tech had our guy run diagnostics on the Adtran Mux in both locations. Both seem to be operating properly. His suggestion was a timing issue. Seems to make sense to me as the connection starts flaking out whenn we put load on the line.

The Adtran tech suggests that the clocking is coming from our Definity phone systems in New Jersey. I always figured verizon would provide clocking for the line, but apparently not (this per Verizon). We are slated for a conference call between the phone system vendors and Adtran tech support for today.

If it's a point-to-point circuit then the Telco does not provide clocking. You should be in good hands with the Adtran folks!

The Adtran folks have been outstanding. I'm not sure if we're any closer to a solution, but I've got to say that I'm really impressed with these guys.

We've bought a couple of TSU120e models on ebay and the Adtran support is helping us get them setup. While I doubt that we've got a hardware issue, at least with the 120e models we can telnet to the devices to check errors remotely. Plus we'll have the piece of mind knowing that we've eliminated one more piece of the puzzle.