Thread Links Date Links
Thread Prev Thread Next Thread Index Date Prev Date Next Date Index

[802SEC] Network status - Tuesday AM

Dear EC members,

I asked Tim Godfrey and Steve Schroedl to provide a status update on the 
network services situation this morning--it is attached below.  The 
objective is make sure you are all appraised of the situation--which is 
critical especially given the problems we have been experiencing.  It 
appears the major problems have been resolved, but I wouldn't be surprised 
if the there continue to be minor issues that pop up.

Both Tim and Steve are ready to address issues, so please let them know of 
any problems you have.



----- Original Message ----- 
From: "Tim Godfrey" <>
To: "Paul Nikolich" <>
Cc: "Steven Schroedl" <>
Sent: Tuesday, July 19, 2005 12:48 PM
Subject: Network status - Tuesday AM

> Paul,
> Here is a recap of the issues and current status of the network:
> When the network was being set up there were problems with the VLAN 
> settings in the hotel network switches. Verilan worked with Core and spent 
> 14 hours on Saturday and Sunday fixing these issues. The VLAN is used to 
> interconnect the 802 network (servers and APs) within the hotel. During 
> setup, there were difficulties getting access to the hotel NOC and 
> facilities which slowed the deployment.
> There were also problem with external bandwidth to the Internet. We also 
> discovered that we were capped at 10Mbps, although we were told by Core we 
> would have a full DS3 (33Mbps sustained).  Core pointed to XO (the ISP) as 
> the problem. By Monday this was still an issue. Although Verilan tried 
> packet shaping, it was still not possible to provide adequate services 
> with only 10Mbps on a Monday.  By the end of the day Monday, XO had solved 
> the problem, and we could see 33Mbps download, but upload was still 
> limited to 12Mbps. We are still investigating this, since DS3 is 
> symmetrical.
> We had problems on Monday in the street level ballrooms due to congestion 
> on the air in the 2.4GHz band. This was due to the loss of  AP settings, 
> which had to be reset because the passwords were not known to Verilan. . 
> The default AP parameters allowed connections at lower rates, which causes 
> congestion collapse, as clients fall back to lower and lower rates and 
> spend more and more time transmitting, exacerbating the problem. By Monday 
> afternoon, the APs had been re-configured to allow only rates above 
> 11Mbps. This should help significantly in the dense ballroom areas.
> Until mid-day on Monday, external Internet access was through a Nomadix 
> router. This caused problems for some users, as it would block all 
> connectivity until they started their web browser and were redirected 
> through a connection page. So if a user started their VPN before their 
> browser, it would not work, and the user would think the network was down. 
> The routing function was moved to a dedicated server by mid-afternoon on 
> Monday, so this should not be an issue any more.
> Verilan was planning to deploy Vivato panels (multiple beam-formed Access 
> Points in a single device) in the ballroom areas, but one was damaged in 
> shipping. By Monday afternoon, additional Cisco APs were deployed in the 
> ballroom area to make up the coverage and bandwidth of the missing 
> Vivatos.
> Verilan feels comfortable that the network is performing properly today. 
> They are still deploying additional APs in the ballroom area in 
> anticipation of the concentration of users for the Wednesday plenary 
> sessions.  If any WG still has problems with connectivity or throughput, 
> please let us know.
> Thanks,
> Tim.

This email is sent from the 802 Executive Committee email reflector.  This list is maintained by Listserv.