Print

[haroonob]

Hello
    I am new to jppf .I have to use this application in Linux is it possible ?.Second my case using this JPPF is one  Java web based application asynchronously    request to JPPF load balancer and load balance just sent this request to any node which is less busy. It does not wait for free node available.

[lolo]

Hello,

Yes JPPF works on Linux. It will work on any platform that supports Java 1.5 or later.
For example I test it regularly on Fedora Core and Ubuntu.

Regarding your load-balancing question, the nodes work synchronously with the server, meaning that a request will remain in the server queue until a node is available.
This is a design choice based on experiments and performance benchmarks for synchronous and asynchronous processing. These benchmarks proved that synchronous processing yields the best performance in every use case.

On the other hand, the JPPF client, which is used to submit requests from your application, allows you to choose whether you want to wait for the response, or fetch it later, so that your application can do other things while the job is executed by JPPF.

I hope this answers your questions.

Sincerely,
-Laurent

[haroonob]

Thanks for replay  
   Its mean that in a real environment in which user is interacting with system this is not use able as my understanding from your answer
       For example we have two nodes and one server .first user request and its process took 15 min to run  and second also join with in second and its process takes 15 min .if third user join after 2 min and it process is only test based and take only 1 min .third user have to wait for freeing up the node .its not process can run with Simultaneously   .
 
     

   

Hello,

Yes JPPF works on Linux. It will work on any platform that supports Java 1.5 or later.
For example I test it regularly on Fedora Core and Ubuntu.

Regarding your load-balancing question, the nodes work synchronously with the server, meaning that a request will remain in the server queue until a node is available.
This is a design choice based on experiments and performance benchmarks for synchronous and asynchronous processing. These benchmarks proved that synchronous processing yields the best performance in every use case.

On the other hand, the JPPF client, which is used to submit requests from your application, allows you to choose whether you want to wait for the response, or fetch it later, so that your application can do other things while the job is executed by JPPF.

I hope this answers your questions.

Sincerely,
-Laurent

[raghu]

Thanks for replay  
   Its mean that in a real environment in which user is interacting with system this is not use able as my understanding from your answer
       For example we have two nodes and one server .first user request and its process took 15 min to run  and second also join with in second and its process takes 15 min .if third user join after 2 min and it process is only test based and take only 1 min .third user have to wait for freeing up the node .its not process can run with Simultaneously   .
 
    

  

Hello,

Yes JPPF works on Linux. It will work on any platform that supports Java 1.5 or later.
For example I test it regularly on Fedora Core and Ubuntu.

Regarding your load-balancing question, the nodes work synchronously with the server, meaning that a request will remain in the server queue until a node is available.
This is a design choice based on experiments and performance benchmarks for synchronous and asynchronous processing. These benchmarks proved that synchronous processing yields the best performance in every use case.

On the other hand, the JPPF client, which is used to submit requests from your application, allows you to choose whether you want to wait for the response, or fetch it later, so that your application can do other things while the job is executed by JPPF.

I hope this answers your questions.

Sincerely,
-Laurent