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

[IDV #MLQ-419928]: ehcache in XmlEncoder

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 Note that the archive is no longer being updated. We provide the archive for reference; many of the answers presented here remain technically correct, even if somewhat outdated.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: [IDV #MLQ-419928]: ehcache in XmlEncoder
  • Date: Mon, 24 Aug 2015 13:10:30 -0600
Don,

More below,

> Julien/Yuan-
> 
> We we are trying to upgrade RAMADDA to use TDS 4.6.  It looks like they
> got rid of the ehcache stuff which is used by XmlEncoder. It looks like
> you've added this to external.jar, but RAMADDA doesn't use that.  So, we
> have a couple of questions:

Indeed we used to get it "for free" via ncidv.jar, but no longer. As a
consequence, it is somewhat annoying to have this heavyweight dependency lurking
around. Moreover, I think ehcache has fallen by the wayside. (Ehcache would also
expose this underlying bug in Java 7 on DHCP networks which was another annoying
problem though that is behind us now.)

> - Why is the cache used in XmlEncoder?  

As I recall, XML deserialization would ask for the same information over and
over again, so pulling info out of cache instead of reconstructing objects from
scratch ameliorated performance somewhat.

> Does it make that much of a performance difference?

Beyond any effects at IDV start up, probably not, but at the time IDV
performance was in the spotlight so we were making a full-court press.

> - Jeff pointed out that if the ehcache.xml file doesn't exist, then the
> static initialization will fail.
> - What is TDS 4.6 using for it's cache now?  

I think they may use Guava Caching (I am trying to confirm this) which is a much
better solution anyway.

> If this is needed, can that be used?
> This is causing an error at RAMADDA startup, so we're trying to avoid this.

In short, this issue is a good candidate for refactoring in the near-term.
 
> Thanks.
> 
> Don
> --
> Don Murray
> NOAA/ESRL/PSD and CU-CIRES
> 303-497-3596
> http://www.esrl.noaa.gov/psd/people/don.murray/
> 
> 


Ticket Details
===================
Ticket ID: MLQ-419928
Department: Support IDV
Priority: Normal
Status: Open