Hi all,
As many of you know, MassLNC issued an RFP in the spring to hire a
consultant to do a performance evaluation for Evergreen. I know we've
been fairly quiet since the RFP was issued, but I'm pleased to announce
that we awarded a contract to OmniTI http://omniti.com/ to perform the
evaluation.
I would like to send along thanks to Rogan Hamby of SC LENDS and Sharon
Herbert of SITKA (along with her great technical staff) who helped
review the proposals and provided feedback to aid us in making a
decision. Also, many thanks to those on the MassLNC team who helped in
the selection process: Dan Pearl (C/W MARS), Jason Stephenson (MVLC),
Martha Driscoll (NOBLE), and Tim Spindler (C/W MARS).
MassLNC has been busy getting a test server set up for the evaluation
(thank you to Thomas Berezansky for his work on this). Once the server
is ready, the evaluation will begin in earnest. I'm sending along
OmniTI's next steps for the evaluation so that the community has an
opportunity to provide feedback on their plans and to ask any questions.
I also asked Pat Norton and Jon Sime from OmniTI to join us at the
developers meeting today in case anyone has feedback to offer there.
CURRENT PROGRESS
As we wait for the test environment to be fully functional, we have
requested both the database logs and Postgresql.conf files from MVLC,
CW/Mars, and Noble. We have begun reviewing both to determine two
things prior to getting into the test environment:
1.) Are there configuration settings that can be recommended in
postgresql.conf
2.) What are the slowest running queries according to the database
logs. As we compile a list of the latter they will allow us to focus on
running those queries in the test environment and observing performance.
NEXT STEPS (ONCE WE HAVE A FULLY OPERATIONAL TEST ENVIRONMENT)
Database
* Execute the worst performing queries and review results and
potential improvements.
* Examine in more detail how Searches are executed and possible
improvements in using T-Search over ILike.
Staff Client
* Put instrumentation in place to observe in more detail where
bottlenecks are occurring
* Determine recommendations based on observance of bottlenecks and
current application architecture
* Once the above strategies are in place in some form, we will be able
to focus on the
specific areas mentioned in the MassLNC RFP. These areas include
potentially migrating to more
recent releases of XULRunner, moving away from XULRunner, locating
memory leaks,
continued use/upgrade of the Dojo toolkit, and the feasibility of
modifying the
architecture to facilitate streaming requests.
Messaging
* Take a detailed review of current messaging architecture
* Determine if alternative technologies can be used to replace all or
part of the existing messaging environment
* If some other technologies can be used, run some tests to compare
using other strategies for messaging in the Evergreen Environment
My hope is that this evaluation will lead to improvements that will
benefit the entire community. Please feel free to send any questions or
feedback our way!
I also want to send thanks along to the Massachusetts Board of Library
Commissioners, which authorized the use of Library Services and
Technology Act grant money to fund a large part of the evaluation.
Kathy
<http://omniti.com/>
--
Kathy Lussier
Project Coordinator
Massachusetts Library Network Cooperative
(508) 343-0128
[email protected]
Twitter: http://www.twitter.com/kmlussier
