So from a CloudTM perspective, this is a part of the adaptive, self-tuning configuration
aspect of things? Or is it auto-provisioning (i.e., adding/removing nodes to meet QoS
parameters)?
- Manik
On 4 Dec 2011, at 13:37, Paolo Romano wrote:
Hi,
one more result from the Cloud-TM project that we thought might be
interesting for the Infinispan community (and possibly OpenShift).
Our last effort is a system for automating elastic scaling for
Infinispan, which we named TAS: Transactional Auto Scaler.
TAS uses a hybrid methodology combining analytical modelling, for
forecasting the effects of data contention, and machine learning
techniques (we experimented both with Radial Basis Functions Aritificial
Neural Networks and Regression Decision Trees), for forecasting the
effects of contention on hardware resources.
Applications of TAS range from on-line self-optimization of
in-production applications, to the automatic generation of QoS/cost
driven elastic scaling policies, and support for what-if analysis on the
scalability of transactional applications.
Results look pretty good, probably even better than we were originally
hoping! ;-)
Cheers,
Paolo
_______________________________________________
infinispan-dev mailing list
infinispan-dev(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/infinispan-dev
--
Manik Surtani
manik(a)jboss.org
twitter.com/maniksurtani
Lead, Infinispan
http://www.infinispan.org