Re: difax problems on IDD ... wondering how Universal?

Paul Ruscher writes:

Haven't seen much traffic here lately, but we have noticed with increasing 
frequency over past few months problems with IDD Difax from Alden.  Many times 
calls to report problems or request backups are futile, as we don't always get 
a response to our phone/email queries, which usually come from our weather 
station manager.

At best, we seem to lose one complete upper air/model cycle per week, and often 
more than that.  It often seems that Alden never gets the maps on their disk 
for backup and so we can never retrieve them.  Difax is dying, I know...but for 
the time being it is still a service that is paid for.  Are satellite users 
experiencing same kind of problems IDD users do?  Is NWS having problems 
getting Difax products to Alden, so that they can't get them out to the Unidata 
sites?  Are other IDD sites experiencing same kinds of data losses?

We'd appreciate some comments from other sites on this one...David Knight's 
post helped prompt this request, and a visit to the IDD stats pages reveal that 
we are not much worse off than other sites.

Paul Ruscher

Up until recently we were getting DIFAX via satellite.  It was very reliable 
and the few times we lost maps were due to problems on our end like a jammed 
printer, etc.

We've been on IDD DIFAX for a month or so now and the service has not been 
great.  Some specific complaints:

1) Reliability: IDD in general has had many problems as of late that the folks 
at Alden and Unidata have been working very hard to solve.  Many of the 
problems such as frequent network outages are well beyond what they can 
control.  Despite this, IDD DIFAX tends to be less reliable then the IDD FOS 
data even though they enter IDD at the same site.  This morning for example 
(while we were without maps for 10 hours) FOS data was getting thru just fine 
while I was not able to LDMping the DIFAX server, thunder.

Part of this problem is likely due to the old slow machine they are running the 
DIFAX IDD on now with old LDM software.  Hopefully this will improve a good bit 
as they most to a new system running LDM5 over the next week or two as I 
understand it.

2) Alden's lack of interest in the service. As I understand it IDD DIFAX customers represent a rather tiny portion of Alden's customer base...and therefore recieve an equivalent portion of their attention.
The last several times I've called their service number to inquire about an 
outage, I've ended up having to explain what IDD DIFAX is to the person I end 
up talking to....  who isn't even aware that they offer such a service..  They 
often start by asking me if our dish is working, etc.  even though I explain 
carefully that we get the data via the Internet.

To be fair, when I've managed to directly contact Mark Franklin or one of the few others there that knows what's going on, they've been very helpful.
One might suspect that the less informed others I've talked to were the same 
people I spoke with several years back who could talk you through major surgery 
on the guts of one of those old wet paper chart recorders.  I was always amazed 
at how well they really knew their stuff....perhaps its just that nobody there 
has bothered to explain the newer technologies to them.

3) While the satellite broadcast maps came in like clockwork, the IDD DIFAX 
maps come in much more at random....Sometimes there will be nothing for a few 
hours and then suddenly a burst of a dozen maps.  They all still get there, but 
if you are looking at the hourly radar summary for example, its nicer to have 
it come in in the same hour it was issued in.

As it was explained to me, the IDD DIFAX maps come from their system by which 
one can call up and request a map be sent to one's fax machine.  This was 
logical for them since that system already stored the maps in G3 fax format.

The problem here is though, that presumably, that system doesn't get the maps 
until they go out on the standard satellite broadcast.  Then it has to crunch 
them for a while to get the in the correct format, etc.  Then finally it has to 
be sent into the IDD system.  Sometimes this takes longer than at other times.

4) The new 2XXX numbering system is based on the product numbers in their telephone fax chart service rather than the traditional "D" numbers we're all familiar with. This would be fine except for the fact there are fewer 2XXX numbers used than the D numbering system had. Several D numbers map into the same 2XXX number. For example....the 9Z Radar Summary is issued as D099 The 10Z Radar summary is issued as D110. In Alden's 2XXX numbering system, they are come out as 2080. (In fact all 24 Radar Summaries, which all have seperate D#'s, end up as 2080.) The same is true for other types of products. This makes it more difficult to automatically identify specific products for special handling such as whether to print or not to print.
I attempt to convert back to D#'s here since our users are more familiar with 
them by using a simple lookup table....the problem is that it didn't work since 
one 2XXX number could be any of more than a dozen D#'s in some cases.  Attempts 
to guess based on time have been unsucessful because of the irregularity at 
which maps are transmitted.  (i.e Is that the 10Z radar summary  coming in or 
just a late 9Z one?)

I'm not arguing that we must go back to D#'s since one set of numbers is as 
good as another, but am rather asking for a finer granularity in distinguishing 
between charts.

5) I've had problems with edges of maps being cut off that I don't recall happeing on the old 
system.  I also am having trouble getting the "wings" or "part2 extensions" on 
the sfc analysis to come out at the same size...though I've taken reasonable care to ensure that my 
printing process is not rescaling them.  (Though I am not ruling this out)


Please understand that I am not trying to just sit and complain.  I understand 
that such things can be complicated.  I've had just as many if not more 
problems that were entirely at our end during our first month on the new system 
while I try to iron out the bugs in our processing system.  And certainly I 
will not blame any one but myself for that.  The above are just some of the 
issues that I see beyond my realm of control...and offer them as constructive 
criticism.

Still, we are paying for the service, so we should expect a certain level of 
comittment on Alden's part to provide it well and reliably.

--------------------------------------------------------
David Wojtowicz, Research Programmer/Systems Manager
Department of Atmospheric Sciences Computer Services
University of Illinois at Urbana-Champaign
email: davidw@xxxxxxxx  phone: (217)333-8390
--------------------------------------------------------