Re: possible solution to performance issues with FTP client on HP3000

Hello John,

James, Does setting the lower bound lower than the initial retransmission
interval accomplish anything?  I was under the impression that the first
retransmission of a packet uses the initial interval, and the intervals
grow longer after that.  Am I mistaken?


[1  ]     Retransmission Interval Lower Bound (Secs)
[2  ]     Initial Retransmission Interval (Secs)

This is the lowest configurable values for these 2 variables (and yes the help documentation for "Retransmission Interval Lower Bound" is wrong... the lowest value is [1] seconds). This is the configuration that I would use in all cases with exception to ***weird*** network's like TCP over satellite or using 3000's as a router and in these cases I would want a network expert involved in devising a timer change as a "workaround".

A feature of the MPE implementation of the "Retransmission Interval Lower Bound" is that the lowest value is [1] second. Note: On some machines it is possible to be less.

**FOR A NEW CONNECTION BEING ESTABLISHED, the "Initial Retransmission Interval" time is used. SYN packets are sent out on the basis of this timer. The first packet is sent out on a connection request and if no response, the second SYN at 2 seconds, the third at +4 seconds, the forth at +8 seconds, the fifth at +16 seconds, the sixth at +32 seconds, the seventh at +64 seconds and the eight at +128 seconds.

I did 8 iterations here because I specified in the configuration the value of:

[8  ]     Maximum Retransmissions per Packet

Note: The number of actual retransmission may actually be bounded by the timer:

[360  ]   Maximum Time to Wait For Remote Response (Sec)

If the sum of time for the timers exceeds this configuration value the retransmission is not performed and the connection attempt is dropped... or after the 8th retransmission times out, the connection attempt is dropped.

----------

**FOR AN EXISTING CONNECTION the TCP packet transmission starts at the "Initial Retransmission Interval" configured value of [2] seconds and then is computed (for the details see Douglass E. Comer's book - Internetworking with TCP/IP [fourth edition - vol 1] - page 226) to be the smoothed round trip time to send a TCP packet and receive an acknowledgement with a minimal value of the "Retransmission Interval Lower Bound (Secs)" configured value of [1] second.

On local LAN links and extended links where the performance of an acknowledged TCP packet is on a smoothed average less than [1] second the "Retransmission Interval Lower Bound" [1] second will be the computed "Retransmission Timer" value. For poor quality links, the "Retransmission Timer" value will increase to be the smoothed round trip time to send a TCP packet and receive an acknowledgement which may be greater than [1] second or [2] seconds or [5] seconds or more...

----------

The BOTTOM LINE answer to your question... is the configured TCP timers are being used for (2) purposes... first for the initial connection establishment (syn-syn/ack-ack) and secondly for ongoing data transfer. The "Initial Retransmission Interval" comes into play on initial connection establishment and is a starting point for initial data transfer. The computed "Retransmission Timer" value is the real timer performing retransmissions once a connection is established, but for very fast and error free networks this timer has a lower bound of the "Retransmission Interval Lower Bound.

I hope this helps...

P.S. Making the "Retransmission Interval Lower Bound" bigger than the "Initial Retransmission Interval" will break MPE's TCP just in case you are curious.

Regards,

James Hofmeister
Email: <first>.<last>@hp.com
Hewlett Packard - Global Solutions Engineering (WTEC)
P.S. My Ideals are my own, not necessarily my employers. 

* To join/leave the list, search archives, change list settings, *
* etc., please visit http://raven.utc.edu/archives/hp3000-l.html *

.

Relevant Pages

  • slow tps on NetWare
    ... databases. ... C Clear Physical Record[Unreassembled Packet] ... [TCP ... Retransmission] R OK ...
    (comp.lang.clarion)
  • Re: Telnet over WAN latency troubleshooting
    ... > we've also tried the old culprit SET PROTO TCP/NODELAY with no ... > with 1200 byte packets I see zero packet loss except for peak time ... was at least one retransmission. ... > I need to find out if there's any way to get telnet to work more ...
    (comp.os.vms)
  • FreeBSD tcp backoff problem
    ... I work on a stack which is derived from FreeBSD. ... but appliance does not receive it ... Since each retransmitted packet is acked, ... it is set to zero when the retransmission timeout happens. ...
    (freebsd-net)
  • Re: Atheros, hardware access layer, collisions
    ... >> I think there is still collision detection happening on the hardware ... I think I have to disable the retransmission of frames ... I've got two computers synchronized to send one packet each to this ... to each that it receives (on the application level, not in the control frame ...
    (freebsd-hackers)
  • Re: Time out Schreiben auf CIFS-Mount
    ... ich glaube, ... This parameter controls the number of times that TCP retransmits an individual data segment before it aborts the connection. ... The retransmission timeout is doubled with each successive retransmission on a connection. ...
    (microsoft.public.de.german.windowsxp.networking)
Loading