|developerWorks editors||Topic: Business intelligence/analytics, database partitioning, test automation, network management, load balancing|
Environment/software: All systems/all software
Level/type: Multiple/information roundup
This is a developerWorks resource roundup on various technology topics -- business analytics, database partitioning, automated testing, and network/load balancing -- that you can use to make your cloud development, deployment, and management experience a smooth one.
Let BI turn your system into AI
developerWorks offers the first three installments in a series that shows you how to use Cognos 8 BI software in the cloud. The Cognos cloud best practices series includes:
- Moving from a single- to a multiple-image topology shows you how to use the Hosts file to manage multiple images, scale out to clusters of different sizes, use private images to create snapshots, and determine where to store files in the cloud.
- Size your architecture for performance and scalability demonstrates the considerations necessary when trying to determine the size of your architecture, including such variables as the composition of your user community, the geographic distribution of users and systems, and the complexity of your applications.
- Enable high availability and disaster recovery provides sample gateway and application server configurations to help you achieve a highly available, reliable system.
Slice, dice, and deal
Database partitioning is an old concept that deals with two concepts key to cloud computing -- data access and data distribution. These two articles provide some new insight on database and database application partitioning:
- Sharding with Hibernate Shards shows how sharding -- partitioning the database at the application rather than table level based on a logical data element -- takes the idea of compacting big data further, allowing databases that don't support regular partitioning to claim a place in the cloud. Amazingly, sharding is something that is unusually best applied later in the process rather than earlier.
- Scale OpenJPA apps with the Slice persistence module demonstrates how single-DB apps can be distributed and horizontally partitioned without changes in the original code or database schema. (Horizontal application partitioning is quite important for performance in the cloud.)
Reuse, not refuse, for your testing code
One benefit the cloud provides for application testing is the ability to rapidly configure a testbed based on existing, predefined components ... the ultimate in reuse. These articles provide insight into ways to develop automated testing components designed to be used over and over again:
- Nuts and bolts to build a Rational Quality Manager reusable test automation system delivers a test automation strategy that shows you how to use RQM as a hub to coordinate the automated test capabilities of other softwares.
- Test web applications with Selenium Remote Control introduces the Selenium framework, a system that lets you build web app tests that can be tested across a range of browsers (browsers being the main interface your customers will have with your cloud application).
Carry that load (efficiently)
With distributed networks, a well-balanced load buys you performance. Gearman improves app performance and load balancing highlights some typical uses of the open source Gearman service that can distribute workloads across the network and speed app and processing requirements.