Infobright, an open source data warehousing company, has announced Infobright Enterprise Edition (IEE) 3.4, which provides performance improvements for analytics involving large volumes of data. The release includes query performance improvements of up to 500%, delete performance improvement of over 1600%, the ability to run queries while data is loading, more efficient workload management, and support for multi-language data. "Our focus is on providing the best query performance possible, but always balancing that with low cost and ease of use," Susan Davis, vice president of marketing, Infobright, tells 5 Minute Briefing.
Infobright's new workload manager now utilizes all available resources including CPUs and memory to optimize the order of data access, and rapidly decompress and read data in parallel with query execution. This advancement has demonstrated query performance increases of 40-500%. The workload management system also provides the ability for multiple concurrent queries to be balanced over all system resources.
In addition, Andrew Flint, product manager, says Infobright has found that many customers are doing essentially a rolling database, meaning that they are keeping their data set current to the last week, or month or 90 days. "What that entails is deleting information off the end as more information is coming in. We have cases where some people are loading up to a terabyte a day and dropping a terabyte a day off the end." In order to support that, Flint says Infobright has improved its delete performance, to the point where it is up to 2000% faster. "We have changed delete to run roughly as fast as our queries do. This is a huge improvement for people who are trying to maintain a snapshot-style database." In addition, the speed of random deletes has been improved as well.
To reduce the wait time for users, large results sets are now streamed to the client, rather than waiting for the full query to execute. This provides faster return of data and will lower the memory footprint for most non-aggregate and aggregate queries. It also benefits ISVs and SaaS providers who embed IEE within their application or service.
Additionally, to ensure faster access to information, IEE 3.4 includes the ability to execute queries while data is still loading. When a load or DML process completes, the new data is immediately available for analytics, enabling business users to load and query concurrently. "Now, queries will continue to run whether there is a load or DML operation in progress, so from a user perspective that is a much faster response time and looks like a huge performance boost."
IEE 3.4 also expands the solution's international capabilities with the ability to load UTF-8 data via the Infobright high-speed loader; support for exporting UTF-8 data; and the ability to convert current ASCII tables to UTF-8.
For more information, visit the Infobright website, or join its open source community.