>Jeremy - even better, what protocl is self correcting ? I need that
>protocol running on my network ASAP !
>
>Nick Payton
Forward error correcting protocols accept addional overhead to
provide enough redundancy to give the receiver a fighting chance to
correct the frame without retransmission. They tend to be used in
radio applications, the extreme case being deep space missions where
the probe doesn't have the power or antenna to do routine
retransmission, and where the speed of light delay is in minutes or
longer.
Another approach to self correction can be seen in such protocols as
SSCOP, which have options for sending the same message over parallel
physical links, and retransmitting only if a frame with a correct
checksum is not received on any link.
While not strictly error correcting, TCP is highly self correcting
with respect to congestion, although there is a continuing evolution
of corrective mechanisms.
>
>-----Original Message-----
>From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]On Behalf Of
>Jeremy Dumoit
>Sent: Thursday, February 08, 2001 8:32 PM
>To: [EMAIL PROTECTED]
>Subject: Dumb question
> I think I'm unclear on some of the protocols here... for what purpose
>would a protocol detect errors, but not correct them? Maybe QoS?
Several reasons.
One, the nature of the application is such that some errors are
tolerable, and it is worse to delay the packet than drop it. Think
packetized voice.
Second, you need to look at the overall protocol stack. If you know
a higher- or lower-layer protocol will retransmit, why bother
duplicating error correction? Think of NFS over RPC over UDP, where
RPC does the retransmission at the record level. Alternatively,
think of UDP over X.25.
Third, the topology is such that it's impractical to retransmit.
Think one-to-many multicasting such as sending weather maps to
thousands of airports. Individual errors are tolerable here, because
weather only changes significantly at 5 or 10 minute intervals (or
longer), and a new copy of the weather map is sent every 30-60
seconds. Statistically, you just need to wait and you will get a
clean copy.
--
"What Problem are you trying to solve?"
***send Cisco questions to the list, so all can benefit -- not
directly to me***
Howard C. Berkowitz [EMAIL PROTECTED]
Technical Director, CertificationZone.com
Senior Mgr. IP Protocols & Algorithms, Core Networks Advanced Technology,
NortelNetworks (for ID only) but Cisco stockholder!
"retired" Certified Cisco Systems Instructor (CID) #93005
_________________________________
FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html
Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]
