With the newest update of its machine-data search engine, Splunk has expanded the user interface in a number of ways so it can be more easily used by business analysts as well as system administrators.
The Splunk software collects and indexes machine data, such as server logs, so they can be analyzed for troubleshooting and usage trends. Although Splunk was originally used mostly for system troubleshooting, more organizations have been using it for discovering usage trends of their systems, which can be used to better customise products and services. National Public Radio, for instance, has used the software for advanced Web analytics.
Revamped user interface
While Hadoop has been getting quite a bit of publicity lately for being a technology for big data-styled analysis, Splunk can also be used to sift through large amounts of data, noted Curt Monash of Monash Research. "Whatever data you can extract from your logs, there is a good chance that Splunk can help," he said.
To help introduce Splunk to a wider array of users, Splunk has revamped the user interface. With the new version, users can rearrange the dashboards directly from dashboards themselves. The previous version required making changes to a text-based XML file. "Everything is editable through dashboards on your desktop, or on your tablet," Mehta said. Users can change around the order of the data charts on the dashboard, or change chart types, for instance, swap out a pie chart for a histogram.
Through the Splunk query language, users can also juxtapose different variables, such as time of day and types of users, to create entirely new charts. Once a user specifies a search through Splunk's query language, that query can be scheduled to run on regular intervals and the results posted in visual form, through charts and other visualisations.
Historical and real-time data together
The software now supports a new type of visualisation, called sparklines. Created by visualisation guru Edward Tufte, Sparklines merge many individual data points into a single illustration that stretches across the screen. "You can not only look at the macro timeline for a search, you can see the granular trending as well," Mehta said. The new version also allows users to fold in historical data with real-time data, expanding the scope of searches.
The software also features a number of improvements on the administrative end. The company claims the new version can work up to 10 times faster, due to a number of reasons. The software has been refactored to allow for many more concurrent users within a single deployment of software, noted Steve Sommer, Splunk chief marketing officer. The software also implements a Bloom Filter, a mathematical technique for filtering out large swaths of irrelevant data for certain types of searches.
A Bloom Filter can "exclude looking at the individual event details and just look at the one data structure, which will allow us to drastically improve search performance," Sommer said.
The software can also support multiple LDAP (Lightweight Directory Access Protocol) implementations, making it accessible across large enterprises with multiple user directories.