All the technical detail, expertise and advice


Celebrus v8 raises the bar

Published: Sunday, 04 January 2015 13:33 by Lincoln Goldspink, Chief Technology Officer
Big Data Analytics

The Celebrus V8 Big Data Engine was a significant leap forwards in the realm of online data capture. Celebrus has always been a 'capture everything' approach for HTML based technologies, but this sometimes under fire from traditional taggers. Why? They have suggested that Celebrus has simply moved the complexity from tagging to server-side configuration. Well, let's take a step back and consider exactly how Celebrus processes and transforms complex clickstream data.

Which is easier, changing website content such as tag definitions encompassing business meaning across multiple brands and touch points, or changing a single server-side piece of configuration?  Because Celebrus is contextualising and processing all of the session data in real-time, session level variables and segments, such as campaign and user type, only need a single server-side configuration to apply across all aspects of the session - rather than being maintained via an in page data layer and placed in every server call where you may wish to use that segmentation. Which is easier to test? Which can be instantaneous? Celebrus.

V8 raising the bar even higher
Celebrus has raised the bar further with V8. Not only does its unique patented data capture technology, which permits dynamic configuration on a page by page basis, provide fast deployments, but server-side configuration used to place business meaning onto the collected event stream now comes with test cases. (V8 Big Data Engine)
The configuration app can record test data in situ from an embedded browser. This browser instruments the target website on-the-fly, using the same data collection agent as per a live deployment. The data collected is sent directly to the configuration app, hence no data collection server need to deployed, nor does an instrumented website exist to form test cases. Cool stuff eh.
Once recorded the test case can be replayed using the current, in-edit configuration local to the configuration app. This is the configuration which isn't as yet submitted to live deployment. Differences between test runs are clearly highlighted, providing side-by-side comparisons of old versus new data based upon current configuration.
Pre-deployment setup and test
What does this mean? Even faster deployments. The ability to configure the system prior to the website being instrumented; even prior to a data collection server being deployed.
Data delivery and data shaping
V8 provides real-time data export and subsequent loading into a variety of structured, semi-structured and non-structured data persistence technologies.  The export is defined by selecting the relevant data items from the event stream and landing them in the appropriate tables, collections or hierarchy, depending upon the target technology. This data shaping capability provides a quick and easy way to tailor the system to meet your needs. There’s a comprehensive schema defined out of the box to use as is, or shape it appropriately.  
What's more, this data shaping is now integrated with test cases too, permitting shaped data tables to be viewed prior to deployment all within the configuration app.
Where next?
Real-time session contexts/states are queued to be included in the test case integration. Once done it will be possible to view the real-time state which will result given a test case.

Fail fast
The earlier problems can be detected, the less time is wasted and the less impact the issue is likely to have. Test cases bring a fail fast approach to the centralised configuration. Failing fast really is the quickest route to success.