Q33. Creating a Parallel Processing Application in Windows Azure?
Cloud computing is greatly suitable for parallel processing. Parallel processing is the execution of application code simultaneously in many machines at the same time. This is usually done to complete a processing intensive task and arrive at the solution quickly.
The first step is to split the processing task into smaller tasks. The tasks are scheduled using the HPC scheduler and application code is simultaneously processed in many machines at the same time. Using this technique of processing, high performance can be achieved and results for complex processing jobs can be arrived at great speed.
In Azure, applications exist as web roles and worker roles and since worker roles execute the business logic, having multiple worker role instances (multiple VMs) facilitates parallel processing. For on-premise parallel processing applications, large cluster of computers are used to get the higher performance, which is very expensive. Comparatively, parallel processing applications in Azure works out less expensive as the compute power is hired only when needed. Many cloud providers are offering server grade processing power for executing parallel processing jobs, on a hire basis. Cloud based parallel is being used by companies for intensive parallel processing jobs.
Azure is able to offer high application performance because it divides an application into separate components and distribute these components on different compute nodes. Consider an example of a job such as credit card processing. The scheduler distributes the job logic on computing nodes and enables quick completion of the job. The HPC job scheduler is equipped to distribute the jobs on on-premise work stations dedicated HPC servers or Azure instances running in the cloud.
The first step is to split the processing task into smaller tasks. The tasks are scheduled using the HPC scheduler and application code is simultaneously processed in many machines at the same time. Using this technique of processing, high performance can be achieved and results for complex processing jobs can be arrived at great speed.
In Azure, applications exist as web roles and worker roles and since worker roles execute the business logic, having multiple worker role instances (multiple VMs) facilitates parallel processing. For on-premise parallel processing applications, large cluster of computers are used to get the higher performance, which is very expensive. Comparatively, parallel processing applications in Azure works out less expensive as the compute power is hired only when needed. Many cloud providers are offering server grade processing power for executing parallel processing jobs, on a hire basis. Cloud based parallel is being used by companies for intensive parallel processing jobs.
Important words:
HPC scheduler
Azure is able to offer high application performance because it divides an application into separate components and distribute these components on different compute nodes. Consider an example of a job such as credit card processing. The scheduler distributes the job logic on computing nodes and enables quick completion of the job. The HPC job scheduler is equipped to distribute the jobs on on-premise work stations dedicated HPC servers or Azure instances running in the cloud.
