[Sis-csi] 2:30 Eastern it is -- IDEA VxWorks info

Keith Hogie Keith.Hogie at gsfc.nasa.gov
Thu Feb 1 16:45:31 EST 2007


Ed,

   I would expect that a frame size of 1024 BITS or 128 bytes should,
at a minimum, support the same frame rate.  However, as I think you
are pointing out, that only results in a throughput of about 1/8 of
the data volume at the same frame rate.  This would drop throughput
but still keep us up in the 10-20 Mbps range.

   So it would seem that an upper rate of 25 Mbps is still OK but
we are starting to push some limits.  As you mention, we still need
more details from Larry to determine if the current limit is based
on bit rate/volume or packets/frames per second.

   I have heard that the LDPC frame sizes being discussed are 1024,
4096, and 16384 bits.  At low data rates like 18/24 Kbps the 1024
bit LDPC is much nicer since it only adds about 1/18 of a second
of latency to collect a frame, code it, and ship it.  Using the
16384 bit LDPC results in 16/18, or almost a second of latency.

Is there a consensus on using a particular size like 1024 bits
and do they plan on using the same size in both directions?

Thanks,

Keith Hogie

Edward Greenberg wrote:
> What would a frame size of 1024 bits do to the performance.  Is the 
> limit bit rate/volume of data unit handling?
> 
> At 3:09 PM -0500 2/1/07, Scott, Keith L. wrote:
>> Wow, that's great info!
>>
>> OK, so those machines top out somewhere between 15k and 30k packets
>> (VCDU frames)/s and can handle at least 132Mbps through.  To a rough
>> order approximation, let's say that any IP QoS (marking, cbq, etc.) we
>> might want to do is about as compute-intensive as whatever VCDU
>> processing is done.  In any case, that kinkd of uplink rate is waayyyy
>> over current envisioned uplink rates, and probably close to the limits
>> of the RF system (as I feared it would be).
>>
>> VxWorks images are determined by the set of features enabled, but I
>> suspect that most run around a few tens of Megabytes, so even a 10Mbps
>> uplink rate should cover a full (base) software reload in under a
>> minute.
>>
>> Since there's almost surely a trade between speed and
>> reliability/correctness, do we foresee any kind of BER requirement we'd
>> like to see on these high-rate uplinks?
>>
>> I think I should work with Greg to see what the next 'ceiling' above
>> 10Mbps might be in terms of coding / framing.  I suspect that once we
>> can do 10Mbps up then (from the baseband processing standpoint, not RF)
>> 15, 20, or even 40 might not be too hard.
>>
>>         --keith
>>
>> -----Original Message-----
>> From: Schneider, Larry [mailto:larry.schneider at nasa.gov]
>> Sent: Monday, January 29, 2007 6:31 PM
>> To: Scott, Keith L.; sis-csi at mailman.ccsds.org
>> Subject: RE: [Sis-csi] 2:30 Eastern it is -- IDEA VxWorks info
>>
>> We know Linux well but are not VxWorks experts. If someone has some
>> commands we could run or tests we could perform within reason we'd be
>> glad to.
>>
>> The VxWorks systems we have are TSI FDPs. They are  about 4 years old.
>> We think they are ~ 300 MHz. We recall
>> discussions about upgrading them to 433 MHz though that never happened.
>>
>> There are a total of 21 of them located at MSFC (6), JSC (3), WSC (6) ,
>> and a couple of support facilities (2). We have four in our lab.
>>
>> The systems process the ISS Ku-band 150 mb/s downlink as well as the
>> S-band downlink.
>>
>> The systems process 15,000 1100 byte VCDU frames / second in and out.
>> We
>> demonstrated the systems can not process 30,000 1100 byte VCDU frames /
>> second in and out though we didn't try to probe any further.
>>
>> We also demonstrated that they can't record and distribute 15,000 1100
>> byte VCDU frames at the same time.
>>
>>
>>
>>
>>
>>
>>
>> -----Original Message-----
>> From: sis-csi-bounces at mailman.ccsds.org
>> [mailto:sis-csi-bounces at mailman.ccsds.org] On Behalf Of Scott, Keith L.
>> Sent: Thursday, January 25, 2007 2:42 PM
>> To: Scott, Keith L.; sis-csi at mailman.ccsds.org
>> Subject: RE: [Sis-csi] 2:30 Eastern it is
>>
>> People I heard on the call:
>>     Keith Scott
>>     Howie
>>     Keith H.
>>     Scott B.
>>     Ed Criscuolo
>>     Will Ivancic
>>     Larry Schneider
>>     Ed Greenberg
>>
>> ** Request from the high-rate uplink WG
>>
>> Current TDRSS uplink rate is 25Mbps (Dave?) There seemeed to be some
>> confusion about what the real rate(s) were:
>>     S-band is 2-4 mbps
>>     Ku ~3Mbps up and down
>>
>> Scott: to estimate the required uplink data rate, find out the maximum
>> rate at which you can forward packets, and use that.  If the first
>> spacecraft on orbit has to serve 3 others or 60 others, doesn't matter,
>> they get what they get but at least we'll be able to get the data up as
>> fast as we can forward it.  For this exercise, assume a dedicated
>> processor for packet handling.  Keith H. notes that several years ago
>> i386 class machines could saturate a 10Mbps Ethernet, so even rad hard
>> things nowadays should be able to do that.
>>
>> What rates can VxWorks forward at over the past N years?
>> VxWorks -- how fast can the network stack forward.
>> Scott and/or flight software people at Goddard may have some VxWorks
>> machines.
>> Larry Schneider has a lot of VxWorks machines to handle the ISS
>> high-rate uplink
>>     VxWorks expert is off this week but email out questions to
>> Larry
>> (to give to
>>     Alex).
>>
>> Ed -- driver for uplink may be HD uplink video (~5mbps) Cap everything
>> by about 30-50Mbps because we don't think uplink radios can do more.
>>
>> Will sent email
>> "What is 'uplink'?"
>>
>>
>>         --keith
>>
>> -----Original Message-----
>> From: sis-csi-bounces at mailman.ccsds.org
>> [mailto:sis-csi-bounces at mailman.ccsds.org] On Behalf Of Scott, Keith L.
>> Sent: Thursday, January 25, 2007 12:35 PM
>> To: sis-csi at mailman.ccsds.org
>> Subject: [Sis-csi] 2:30 Eastern it is
>>
>> Ok, we're on for 11:30 pacific, 1:30 central, 2:30 eastern on
>> Thursdays, starting today.
>>
>> Phone Number:
>>     703 983 6338 (x31550) in Washington
>>     781-271-6338 (x16338) in Bedford
>>     866-648-7367 (866-MITRE-MP) Toll Free
>>
>> Meeting ID: 55555
>>
>> Today's Agenda:
>>     Recap from Colorado Springs Meeting
>>     Request for requirements info from High-Rate Uplink WG
>>     Red Book 1 Items:
>>         Collect information and distribute
>>             GPM (Jane)
>>             CANDOS (Dave/Keith H.)
>>             SNIS (Dave)
>>             Shuttle&ISS (Larry)
>>             Surrey satellite (Loyd/Will)
>>         Mash these against requirements
>>         From the 'assumptions' slide, list possible solutions
>> to each of the requirements
>>         Downselect possible solutions
>>         Assign people to write text
>>         Publish book
>>
>>         --keith
>>
>> -----Original Message-----
>> From: sis-csi-bounces at mailman.ccsds.org
>> [mailto:sis-csi-bounces at mailman.ccsds.org] On Behalf Of Scott, Keith L.
>> Sent: Thursday, January 25, 2007 9:34 AM
>> To: sis-csi at mailman.ccsds.org
>> Cc: Greg.J.Kazz at jpl.nasa.gov
>> Subject: [Sis-csi] Telecon scheduling and request for requirements
>>
>> ========== Telecon scheduling
>>
>> I apologize for this but we're going to need to reschedule our weekly
>> telecons, starting with today's.  Hopefully this will be the last time
>> we'll have to do this for a while.  I'll start off by proposing one
>> hour earlier on Thursdays (11:30 pacific, 1:30 central, 2:30 eastern).
>> Please let me know if this works for you and, if not, what would.  If
>> we can't come to consensus quickly we'll wave off this week's telecon,
>> but please see below on a request for requirements from the High Rate
>> Uplink WG in SLS; I'd like to hash this out over email and get them an
>> answer by next week.
>>
>> ========== Request for Requirements from High-Rate Uplink WG
>>
>> Following up from the Colorado Springs meetings, we've received a
>> request for requirements from the High Rate Uplink WG (CWE at
>> http://public.ccsds.org/sites/cwe/sls-hru/default.aspx?RootFolder=%2fsi
>> tes%2fcwe%2fsls%2dhru%2fPublic%2fCharters&View=%7bED7A93D1%2d1BB1%2d403
>> 3%2d8432%2dD7382F71947D%7d).  In particular they want to know what
>> kinds of uplink rates the cislunar architecture would require.  I think
>> that our DOWNLINK requirements should at least cover current shuttle
>> and station rates of ~150Mbps, but I don't think we require full
>> symmetry on the uplink.
>>
>> The largest driver for the uplink rate might be crew accommodation of
>> live HDTV events, which would be on the order of 5-10Mbps, if they were
>> actually provided.  If we had that then we should be able to cover most
>> or all of the operational uplink traffic in the margin!  :)  Just as
>> another point of reference, 150Mbps of TCP downlink would require
>> ~2Mbps of TCP ACK traffic.  We also need to remember that in a
>> networked world such as we've proposed, one uplink may be serving
>> several spacecraft.
>>
>> In any case, we need to provide them with a number that:
>>     1) Is large enough to cover our requirements for a reasonable
>> amount of time
>>     2) Is not so large that they roll on the floor laughing at us
>>
>> Regardless of the rate requirement, RFC3819 (Advice for Internet
>> Subnetwork Designers) and maybe rfcs RFC3155 (End-to-end Performance
>> Implications of Links with Errors) and RFC3366 (Advice to link
>> designers on link Automatic Repeat reQuest (ARQ)) seem relevant.
>>
>>         --keith
>>
>> _______________________________________________
>> Sis-CSI mailing list
>> Sis-CSI at mailman.ccsds.org
>> http://mailman.ccsds.org/cgi-bin/mailman/listinfo/sis-csi
>>
>> _______________________________________________
>> Sis-CSI mailing list
>> Sis-CSI at mailman.ccsds.org
>> http://mailman.ccsds.org/cgi-bin/mailman/listinfo/sis-csi
>>
>> _______________________________________________
>> Sis-CSI mailing list
>> Sis-CSI at mailman.ccsds.org
>> http://mailman.ccsds.org/cgi-bin/mailman/listinfo/sis-csi
>>
>> _______________________________________________
>> Sis-CSI mailing list
>> Sis-CSI at mailman.ccsds.org
>> http://mailman.ccsds.org/cgi-bin/mailman/listinfo/sis-csi
> 
> 
> _______________________________________________
> Sis-CSI mailing list
> Sis-CSI at mailman.ccsds.org
> http://mailman.ccsds.org/cgi-bin/mailman/listinfo/sis-csi


-- 
----------------------------------------------------------------------
   Keith Hogie                   e-mail: Keith.Hogie at gsfc.nasa.gov
   Computer Sciences Corp.       office: 301-794-2999  fax: 301-794-9480
   7700 Hubble Dr.
   Lanham-Seabrook, MD 20706  USA        301-286-3203 @ NASA/Goddard
----------------------------------------------------------------------



More information about the Sis-CSI mailing list