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

Re: 20150516: Nexrad 2: idd.unidata.ucar.edu Access

"The other thing to consider is to split single feed requests for high volume 
feeds into multiple, mutually exclusive feed requests whise sum is the full set 
of products.
Splitting feed requests helps to get around TCP backoff induced latencies."

Could you give us an example of how we would do this in the config?

________________________________________
From: Unidata User Support <address@hidden>
Sent: Tuesday, May 19, 2015 1:56 AM
To: James Olson
Cc: Hunter Goosmann; address@hidden
Subject: 20150516: Nexrad 2: idd.unidata.ucar.edu Access

Hi James,

re:
>We are just curious as to what the difference is for your server to receive=
> LDM data, from what we believe to understand as being similar upstreams of=
> borth the ROC Node and TOC Node, and our connection. I interpret in your p=

There should be no difference.

re:
>revious email you were implying you were seeing high latency at both thos N=
>odes. Am I correct in this interpretation?

We were, yes.

re:
>We've heard little back from our Contacts with NOAA and are a bit weary of =
>attempting to move our primary feeds back over to TOC and ROC. Would you ha=
>ve any recommendations on how we might be able to test things on our end?

You should always feed from the source for data when possible.  You should also 
feed from one or more redundand
sources for all of your data feeds if you can.  The redundancy should help to 
make your feeds as reliable as possible and perhaps reduce latencies.  The 
other thing to consider is to split single feed requests for high volume feeds 
into multiple, mutually exclusive feed requests whise sum is the full set of 
products.
Splitting feed requests helps to get around TCP backoff induced latencies.

Cheers,

Tom Yoksas
Unidata User Support
>
>-----Original Message-----
>From: Unidata User Support [mailto:address@hidden]=20
>Sent: Saturday, May 16, 2015 6:45 PM
>To: James Olson
>Cc: Hunter Goosmann; address@hidden
>Subject: 20150516: Nexrad 2: idd.unidata.ucar.edu Access
>
>
>Hi James,
>
>re:
>>Hi, I'm with the ERC (Education and Research Consortium of the Western=20
>>Caro=3D
>>linas) We're one of the top tier level 2 providers. We're currently=20
>>have dr=3D astic latency issues with both of our upstream. Anywhere from=20
>>10-20 minute =3D latency of data on one and 40 minute latency of data on=20
>>the other. Unfortun=3D ately there's not much we can do about the=20
>>upstream issues until Monday mor=3D ning. It was suggested to us that we=20
>>may have access to the idd.unidata.ica=3D r.edu ldm server that could be=20
>>used temporarily until we can speak to the p=3D eople incharge of our=20
>>upstream ldm servers on Monday. It appears we have ar=3D e allowed to=20
>>pull from the server, but just want to make sure it's alright =3D we do=20
>>so. My Boss Hunter Goosmann said if any sort of payment is required t=3D=20
>>o do this just speak with him. It's extremely urgent we get our latency=20
>>iss=3D ues resolved as we have a number of customers currently effected by=
> this is=3D sue as well.
>
>nws-ldm[12].hpcc.ercbroadband.org are both ALLOWed to REQUEST NEXRAD2 from =
>idd.unidata.ucar.edu AND are both already making REQUESTs for this and othe=
>r feeds.  This should indicate that the latecnies you are experiencing are =
>either caused by the connection(s) to the NWS top level sites, or by networ=
>k issues close to your location.
>
>Latencies we are seeing for the times of your high latencies strongly sugge=
>st that the problem is at or close to the NWS top level source machines.
>
>I hope that this helps...
>
>Cheers,
>
>Tom Yoksas
>Unidata User Support
>--
>********************************************************************** <
>Unidata User Support                              UCAR Unidata Program <
>(303)497-8643                                            P.O. Box 3000 <
>address@hidden                             Boulder, CO 80307 <
>---------------------------------------------------------------------- <
>Unidata Web Support               http://www.unidata.ucar.edu/support  <
>----------------------------------------------------------------------
>< NOTE: All email exchanges with Unidata User Support are recorded in the U=
>nidata inquiry tracking system and then made publicly available through the=
> web.  If you do not want to have your interactions made available in this =
>way, you must let us know in each email you send to us.
>
--
********************************************************************** <
Unidata User Support                              UCAR Unidata Program <
(303)497-8643                                            P.O. Box 3000 <
address@hidden                             Boulder, CO 80307 <
---------------------------------------------------------------------- <
Unidata Web Support               http://www.unidata.ucar.edu/support  <
---------------------------------------------------------------------- <
NOTE: All email exchanges with Unidata User Support are recorded in the
Unidata inquiry tracking system and then made publicly available
through the web.  If you do not want to have your interactions made
available in this way, you must let us know in each email you send to us.