CloudTran Product Overview

Data Processing for Web Applications

We're all familiar with load balancers that scale-out the application layer in a web architecture. But look a little lower in the stack, and you'll find that many applications lack the ability for scaling-out at the data layer. It's not uncommon to find web applications today funneling data processing to a single database entry point, which will eventually result in performance bottlenecks as workloads increase. For applications to truly work at today's web-scale and to take advantage of all computing power on each server, an architecture must include scale-out capabilities at both the application layer and the data layer.

Why You Need CloudTran

Built for Scale

CloudTran is the industry's only solution for managing the complete data layer in a distributed computing environment. We integrate in-memory data grids with back-end data stores and provide ACID transactionality across the entire architecture. CloudTran enables applications like online gaming, mobile payment processing, content monetization, and even mainstream OLTP to leverage all the value of affordable, commodity hardware to provide fast, scalable applications.

CloudTran provides an important capability on top of what's provided by data grid and database vendors. Using CloudTran, you get an out-of-the-box persistence layer that automatically writes data in the grid to back-end data stores. All writes are done asynchronously and in the same transactional context as the data grid. This ensures data consistency and reliability with no special effort on the part of your developer. In addition, CloudTran manages sharded data stores. This provides fast, scalable performance as user loads increase resulting in consistent response times no matter the load.

The CloudTran Architecture
  • CloudTran Application Interface offers developers a choice of ways for working with CloudTran. One method is to use Java Persistence API (JPA), which provides an industry-standard way of accessing all CloudTran features. CloudTran also supports native data grid APIs such as the Oracle Coherence API for compatibility with applications written to this interface. And for a simplified development environment, the CloudTran ORM generates code for mapping all data access to Plain Old Java Object (POJO) for an easy and consistent programming metaphor.
  • CloudTran Transaction Manager provides rock-solid ACID transactions across grid nodes, between the grid and persistent data stores, and across distributed data stores. By unbundling transaction processing from the database layer and acknowledging the data grid as the system of record, applications can run at grid speed with asynchronous write-behinds to disk. The CloudTran Transaction Manager also handles "lights-out" recovery of grid nodes and guaranteed persistence of all data to disk.
  • CloudTran Sharding and Storage Manager handles the mapping of data to specific data storage formats as well as manages data sharding for scale-out at the data store level. All sharding is transparent to the application since CloudTran handles the scatter and gather of data.

CloudTran Architecture

Click to enlarge

© Copyright 2012 CloudTran, Inc.| All Rights Reserved.