On Feb 5, 2014, at 2:19 PM, Sachith Withana <[email protected]> wrote:

> Hi all,
> 
> After the experiment creation and submission, wouldn't it be better to keep 
> the basic experiment information in the gateway itself?  such as 
> experiment_name, ID, Job creation time ...etc.
> 
> If not, then every time the user wants the information for a particular 
> Experiment, the gateway would have to talk to the Server ( Airavata Server in 
> my case) .

I think there will be a bit of both use cases. Gateways wanting to have 
everything and second set of use case a slick portal. I do not have a string 
opinion on this.

> 
> It takes away the delay of the communication as well as the unnecessary load 
> on the server. 

I do not think these will be real concerns. Thrift if supposed to be “high 
performance” api. Think about Ever Note note syncing and the number of 
communications which happens when you write a note synced across multiple 
devices in real time. Secondly on the load on the server, thats should also not 
be a issue. When we hit the limit, we have to load balance. More data is not an 
issue as well, we are in big data era, when we put it in perspective anything 
we talk here will be small.

Suresh

> 
> Any suggestions?
> 
> -- 
> Thanks,
> Sachith Withana
> 

Reply via email to