Streambase on EC2 - demo

After installing Streambase, I went through the getting started guide, it was good, perhaps a version for the Developer Edition might have been useful.

Again there were some hardcoded path issues with the /etc/init.d scripts, looking for the binaries in /usr/bin rather than a passed location. I will add this to the wiki HOW TO document as well.

I had to add some symbolic links to get Streambase to attempt to start and then it failed as the Developer edition is not allowed to be run as a separate server.

Not phased, I moved on using the Creating a Clustered Application documentation and got the cluster management node and processing node configured and running, just to make I understood that process.

At that point, I installed the Streambase Studio onto my Windows XP box and had a look at the various demos available.

One main difference is the way data is treated by Streambase and StreamSQL. The data is manipulated as it arrives rather than storing and then analyzing. So various intermediate steps where data is being tranformed or sliced and diced can be discarded.
Of course various stages can also been stored and also used as a feed into another process.

One of the nice things is that Streambase Studio is built on Eclipse, making the flow of the data visually apparent.

I might drop the Streambase team an email and ask if I can get access to the Enterprise Edition to at least test running the server on EC2 properly as a stand alone server.

So next steps are to go back to Bizgres and Greenplum and run through the clickstream demo and review the clustering/master slave node options with that product.

stay tuned...

Paul