Grid 2.0 – The Next Step for Real Application Clusters

Let me be clear, I am not a RAC guy, but while using the schedule builder this year, one presentation had a definite hint to the upcoming 11gR2 release. Bob Thorne, Director of Product Management (Oracle) gave an insight on Monday, into the new Grid 2.0, via his presentation called “Oracle
Grid Computing 2.0: A Preview“:

Grid 2.0 will include a new grid infrastructure that dramatically extends the
benefits of Oracle Grid Computing. This session covers this new infrastructure,
which will greatly simplify the management of a cluster, propelling grid
computing to a mainstream architecture. New quality-of-service capabilities will
make meeting performance objectives in a shared infrastructure environment easy
and automatic. The session also introduces exciting new storage management
features that extend the benefits of Automatic Storage Management to all files.

His presentation gave insight in the upcoming new RAC infrastructure which had the following characteristics (as far as I can remember):

 1. Dynamic infrastructure

o Automatically change to meet defined certified level objective based on policy (policy based automation)
o Fast provisioning and de-provisioning of resources as needed

2. End to end optimization

o Sharing and optimization of resources across the entire stack (application sever – database – storage)
o Enabling more applications and workloads around the entire grid

3. Always online

o “Always online” means more than just protection from failures
o Zero unplanned and planned downtime
o Online patching, upgrades, application changes
o Real time system health monitoring

a) proactive rejuvenation of application components
b) root cause analysis after eviction of problem nodes

4. Standard deployment

o Grid 2.0 for everything: database, application server, applications, webservers and storage

5. Storage for all

o Datafiles, Normal files, etc.
o A new file system.

So in short, “Grid 2.0” is the next logical step for RAC. Automating the whole stack, from front-end to storage, based on policies that automatically redefines it services were needed in the architecture. Based on a new storage files system, it can now be used for not only database purposes, but also for normal files like the Oracle software. Mr. Thorne also showed a picture of a cluster problem diagnosis tool, “Cluster Monitor V1.01”, that contains all info needed accros the stack to detect problems. This will be made available shorrtly via OTN. Also Oracle 11gR2 was briefly mentioned, which most of this functionality will contain, although the presentation started with the standard disclaimer about future implementations. The release date for Oracle 11gR2 was refered by the statement that “it will probably be released somewhere next year”. So in short: maybe…

😎