Guest

Cisco 600 Series DSL Customer Premises Equipment Routers

Field Notice: *Expired* FN - 9391 - Cisco 6xx - Low FTP Throughput


Revised May 3, 2007

July 6, 2000

NOTICE:

THIS FIELD NOTICE HAS BEEN ARCHIVED AND IS NO LONGER MAINTAINED OR UPDATED BY CISCO.

THIS FIELD NOTICE IS PROVIDED ON AN "AS IS" BASIS AND DOES NOT IMPLY ANY KIND OF GUARANTEE, WARRANTY OR SUPPORT. USE OF THE INFORMATION ON THIS FIELD NOTICE OR MATERIALS LINKED FROM THIS FIELD NOTICE IS AT YOUR OWN RISK. CISCO RESERVES THE RIGHT TO CHANGE OR UPDATE THIS FIELD NOTICE AT ANY TIME.


Products Affected

Product

Comments

Cisco 677

 

Cisco 678

DMT mode only

Cisco 627

 

Problem Description

File transfer performance rates may not meet expectations.

Background

There is an issue with some of the DMT based Cisco 6xx CPEs in that "reliable" file transfers may not occur at the expected transfer rate. "Reliable" file transfers are those which use a protocol having acknowledgement messages. Performance below expectations is most likely to manifiest itself at high downstream trained rates, particularily above 4096 kbps. For example, at the maximum downstream trained rate, the file transfer rate for an FTP transfer may be no higher than 70% of the trained rate. The Cisco 627, 677 and 678 (when running DMT) are the CPE types that are most likely to experience this issue.

To follow the bug ID link below and see detailed bug information, you must be a registered user and you must be logged in.

View bug details for CSCdp23833.

Problem Symptoms

The CPE trains at the expected connection rate, but FTP file transfers and the operation of other "reliable" protocols are slower than expected. The transfer rate for "unreliable" protocols such as TFTP occur at or near the trained rate, and do not appear to be an issue.

Workaround/Solution

There are several suggestions to try to improve performance. Note that some of these items may have drawbacks and may not be useful in all situations.

  1. Make sure that you are running CBOS 2.3.0.053 or higher.

  2. Increase TCP window size, particularily if using Microsoft Windows 95/98 or NT client. Refer to Microsoft's web site at "www.microsoft.com" for information on caveats regarding modification of TCP window size. For Windows 95/98, search for "DefaultRcvWindow delay" and for NT search for "TcpWindowSize high delay"

  3. Utilize traffic shaping. Set the UBR on the PVC of the terminating router to be equal to the trained rate of the CPE.

    Note: This option should especially helpful when the CPE shows excessive CRC errors because of a large TCP window size.

  4. Decrease interleave delay. This parameter is configured on the DSLAM, and there is a direct relationship between a smaller interleave delay and increased performance. Because this parameter defaults to max (for "dirty" line conditions), it may be possible to configure it to a shorter value and improve throughput. However, it is recommended that changes to the interleave delay be done in steps to see how the performance is affected. Too short of a delay can actually hurt performance on "dirty" lines.

For More Information

If you require further assistance, or if you have any further questions regarding this field notice, please contact the Cisco Systems Technical Assistance Center (TAC) by one of the following methods:

Receive Email Notification For New Field Notices

Product Alert Tool - Set up a profile to receive email updates about reliability, safety, network security, and end-of-sale issues for the Cisco products you specify.