Welcome to the new Gigaspaces XAP forum. To recover your account, please follow these instructions.

Ask Your Question

Is Gigaspace suitable to batch applications ?

I have heard about many cases using a space or grid based architecture to support online applications, such as online trading, Web application. I was puzzle whether the architecture can be applied to large scale batch applications.

I have a number of clients which are running large scale batch applications on legacy environment (mid-range, mainframe). Unlike online application which demand very high response time (or low latency), what our client really care is throughput. The highest amount of transaction volume I'm taking about is something like 10,000+ transactions per second (business transactions, not just database commit) and their system need to process transaction at this level for hours.

Currently our client is running their applications on a legacy platform using index file based storage. It is a pretty big box and they turn off all transaction integrity control to reduce overhead (so, if there is anything went wrong, they have o re-run the batch again).

One of the way to better scale the system is to partition the data into multiple independent data sets so that the application can be run on multiple nodes independently. However, unlike online application which each transaction request is usually independent and self contained, it is not possible to completely partition the database into multiple datasets independently. For example, if we partition the data using customer, well, while transactions from different customers are independent from each other, they all update the same set of products. In other words, we can have the customer table partition across multiple nodes, but the product table must be shared across all these nodes (no matter whether it is resided in physical database or IMDG).

My question is, is solution like GigaSpace a right approach to solve the problem of my client ?

{quote}This thread was imported from the previous forum. For your reference, the original is [available here|http://forum.openspaces.org/thread.jspa?threadID=2502]{quote}

asked 2008-07-31 03:42:42 -0600

rickywck gravatar image

updated 2013-08-08 09:52:00 -0600

jaissefsfex gravatar image
edit retag flag offensive close merge delete

1 Answer

Sort by ยป oldest newest most voted

There are several customers using GigaSpaces for similar type of applications. Here are the product features that are the main added values in such scenario:
- large amount of data storage in memory - reduce the need to access the data reside on disk. Faster data retrieval.
- distributed data support - parallel query support across multiple space partitions
- in memory transaction support - avoid the need to restart the whole process from scratch. Preserve data consistency and provide atomicy.
- execution api framework - remoting api allow u to invoke business logic across multiple partitions in transparent using map reduce interfaces.
- database integration - ability to load data from your database in transparent manner
- async persistency service - reliable persistency mechanism allowing u to remove the database out of the critical path of the transaction having data sent to the database as part of background process.
- interoperability - ability for cpp , java and .net to participate the batch process.
- integration with enterprise grid products - we are working with all leading enterprise grid vendors allowing u to increase GigaSpaces grid size on the fly.
- short learning curve and developer tools - u will find web casts , quick start guide , monitoring and deploy tools and ready made examples for batch processing scenarios.



answered 2008-08-02 04:57:49 -0600

shay hassidim gravatar image
edit flag offensive delete link more

Your Answer

Please start posting anonymously - your entry will be published after you log in or create a new account.

Add Answer

Question Tools

1 follower


Asked: 2008-07-31 03:42:42 -0600

Seen: 90 times

Last updated: Aug 02 '08