[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[LDM #XXW-990118]: LDM - Not your average latency problem

This archive contains answers to questions sent to Unidata support through mid-2025. Note that the archive is no longer being updated. We provide the archive for reference; many of the answers presented here remain technically correct, even if somewhat outdated. For the most up-to-date information on the use of NSF Unidata software and data services, please consult the Software Documentation first.

  • Subject: [LDM #XXW-990118]: LDM - Not your average latency problem
  • Date: Fri, 20 Oct 2006 15:45:49 -0600
Mike,

> OK , now you're a believer....check out the latencies, they are starting to
> increase.

Yes, the latencies increased.

> This is what I predicted in my email yesterday. Insane latencies
> on small volumes.......so, it must be a router choking somewhere......or? I
> wonder if some kind of limit kicks in on their ISP when there are too many
> connections. Or, could it be something in the LDM?

I don't know the cause of the problem.  The only thing I do know is that the 
LDM isn't the cause (note that we have no problems receiving the split feeds 
from Rossby).  I suspect that Sonomatech's network is messed-up, somehow.

> And, if we let this go long enough, it would get worse and the FNEXRAD or
> UNIWISC feed might just disappear for long stretches.
> 
> We can let it go until this evening like it is, but then I should probably
> combine the feeds and get it stable....as I'll be gone this weekend. I'll
> probably add back in the entire NIMAGE feed, and deal with 300-600
> latencies at times. STI loved that, it's better than 1 hour latencies
> and/or no data.

Reconsolidating the feeds sounds good.  Go for it.

> Or, if I do want to combine feeds, AND have a subset of NIMAGE data...say
> "WEST CONUS and EAST CONUS" how might I form the "very ugly" request line?

The problem is that the pattern would have to request all the IDS|DDPLUS data, 
plus all the UNIWISC data, plus all the FNEXRAD data, but only the NIMAGE data 
of interest.  I don't know enough about the product-identifiers to even begin 
that regular-expression.

A workaround would be to modify Rossby's ALLOW entries to take advantage of the 
"OK" pattern, e.g.,

ALLOW IDS|DDPLUS|FNEXRAD|UNIWISC ^66\.224\.54\.160$
ALLOW NIMAGE ^66\.224\.54\.160$ <<east or west conus pattern>>

where <<east or west conus pattern>> is an extended reqular-expression that 
selects only the
east or west CONUS data-products of the NIMAGE feed.  Sonomatec's LDM could 
request a pattern of ".*", but would only get the east or west CONUS images.

Regards,
Steve Emmerson

Ticket Details
===================
Ticket ID: XXW-990118
Department: Support LDM
Priority: Normal
Status: Closed