I attended HP Discover in Vienna with some sense of wondering what it might bring. It turned out to be the old HP back, but with a clearer understanding of their core strengths and a new energy and focus on how they are delivering them. As with all such events there was a number of bloggers present providing good coverage of the event overall including an official HP blog.

But that’s enough publicity for the event and onto the topic that interested me, and my colleague Manuel Sevilla, who is the CTO for Capgemini Business Information Management, and who is right in the middle of the whole topic of big data. The term ‘big’ might also be taken to apply to the size of the topic itself and indeed to the amount being written on it! Manuel has written some interesting stuff on this in the Capgemini Capping IT Off blog which includes coverage of SAP HANA and Oracle Exadata, so sitting together through the presentation of the HP approach and products to compare notes was really very useful.

We were both quite thoughtful on why and how HP plans to merge Autonomy and Vertica on one side and the performance and power of the new HP storage and back-up products on the other side. We both felt it could be a game-changer and I will now steal a line from the excellent presentation given by Dr Mike Lynch, the founder of Autonomy, which also provides the title for this post. He reckoned that the ‘T’ standing for Technology in IT had changed three times in terms of the technology in use, but that the ‘I’ for Information had not changed at all, until now. His point was that we are still visualizing big data as an extension of the machine-driven data model, just thinking there is more of it, whereas in fact the change will include substantial amounts of human-centric media data.

He illustrated this point by showing how media-based information would be stitched together in response to recognition of an image and it was pretty impressive demo stuff for such an event, but it was serious in its point too. We recognize that this most useful information is all unstructured, so is not suited to a relational database, but is Hadoop making enough of an accommodation for what it means? Certainly it scales, but with a not-impressive performance in delivery, and human eyes are very sensitive to obvious latencies in responses. So his insights into big data and human use of information, meaning that the data is not just unstructured but is actually made up of all different media types, was excellent, but his demo really did prove that an ‘answer’ could and would be an integration of media together.

At this point do take the time to watch the demo as it really does make you think differently on how things will be, and of course in its own right it is an example of ‘media’ being the best way for a human being to understand a complex topic. So it’s well worth watching this, and when you do try to work out how much time and text it would take to describe this in writing.

It’s an interesting combination that you need the ability to search unstructured data for your unobvious alignments that deliver real insights, hence HP’s acquisition of Autonomy for this rather than to go and compete with existing players in the Business Intelligence market. But you also need extremely quick capabilities to deliver from storage, which brings in the new hardware. However, before I go there I should point out that not everything will be unstructured and it’s probably not going to work to assume that you can separate the activities of use of structured data and unstructured media. That brings in the merging of Autonomy with HP Vertica and its online analytics of structured data, so the HP plan to merge this with Autonomy looks like a real ‘differentiator’ in this all-important big data segment of the market.

For me it made sense of the HP acquisition and for Manuel it excited him greatly with the capabilities!