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 Note that the archive is no longer being updated. We provide the archive for reference; many of the answers presented here remain technically correct, even if somewhat outdated.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.
Robert, >Date: Fri, 13 Jun 2003 10:45:45 -0500 >From: Robert Leche <address@hidden> >Organization: Louisiana State University >To: Steve Emmerson <address@hidden> >Subject: Re: 20030612: LDM6.0.13 Binary for SunOs 5.7 The above message contained the following: > 1) Looks like none of our ntp systems are working (talking to each other) > at the moment. I am working on this. I'l let you know when this is > resolved.<br> OK. > 2) The Server "Hurricane" has hardware problems and as such you will see > it fail. Again I am work ing on it.<br> OK. > 3) This problably fits in the stupid question catogory, but... <br> > Concerning LDM's ldmd.conf > configuration. If I have a LDM IDD system > that is feeding down stream sites. And if in my ldmd.conf I am requesting > data types 1,2, and 3. And one down stream users is requesting data types > 4,5, and 6. And another down stream user is requesting 7,8, and 9. Would > I need to request all of this as: request 1|2|3|4|5|6|7|8|9 ".*" > idd.upstream.feeder.edu The union of requests to upstream sites would need to encompass the union of all the feedtypes requested from you. Whether or not you request all the relevant feedtypes in a single request to an upstream LDM depends on your particular situation. In theory, it is possible to make a single request for all the data. In practice, however, the regular-expressions for the feedtypes might differ, in which case it makes sense to separate them into distinct requests. Also, it turns out that multiple TCP connections usually have better performance than a single TCP connection -- even though the same amount of data is being transferred -- so separating the upstream requests can decrease product latencies. The tradeoffs are between product-latency, the number of child LDM processes that your system can handle, and ease-of-maintenance of the LDM configuration-file. Because each situation is different, there is no hard rule. Generally, I write the configuration-file for ease-of-maintenance and then closely monitor the LDM system's performance for problems. Regards, Steve Emmerson