HP’s enterprise play

The internet is ablaze with posts and articles on HPs abandonment of WebOS and the PC business. I don’t have a lot to add to this discussion (It’s a pity, I was considering a Pre3 device, blah blah). I am personally more interested in the last bit of their announcement that talked about the 10.3bn$ acquisition of Autonomy. If we add this to HP’s acquisition of Vertica just 6 months ago and the (supposedly )failed attempt to acquire Tibco we can see HP is making a  play for the whole unstructured/big-data/analytics field.

Vertica is  a columnar database optimized for analytics. In a nut shell, columnar databases allow fast aggregation of  data as well as holding a lot of columns per row – both traits are helpful when trying to report on data (but vey wasteful when you try to do OLTP operations). Autonomy adds to that path analysis and related algorithms to provide insight from stored data. Add that to HP’s hardware capabilities and you can get a high-performance data appliances that can handle large amounts of data efficiently.

HP is not alone in this trend. We see similar moves by EMC which acquired Greenplum followed by  unified GreenPlum/Hadoop  SW-HW solution, IBM acquiring Netezza and  rolling  its own Hadoop version, heck, even Teradata recently acquired Aster Data to bolster its analytics offering. It’s almost like we’re back in the 1950’s with companies creating dedicated appliances instead of general purpose computers.

While Teradata, an NCR spinoff, pioneered this trend, I think Oracle’s introduction of Exadata popularized this trend. Oracle which is probably the leader could (and still can) be sold with any hardware. All of a sudden, with the acquisition of Sun, Oracle started to keep the high-end customers to itself. This was especially a blow to HP considering the first incarnation of exadata ran on HP’s hardware. Anyway, as we saw above, the big vendors* are rushing to build their own “complete software and hardware suits” and HP, it seems, is doing the same. By talking about dropping its PC business HP is trying to pull an IBM (and its Lenovo move) and focus on Enterprise solutions in general and the hot analytics market in particular

Meanwhile I am reading that HP servers are having trouble handling the demand for their tablet dumping – I guess that’s also bad press to the cloud message of HP. I hope their analytics message and solutions will fare better

By the way, for now this is only interesting for larger corporations which are willing to spend the big bucks involved  (e.g. Vertica was priced at 100K$ per terabyte last time I checked). Smaller companies esp. SaaS provider, will probably still solve their big data problems with Hadoop, Cassandra and the like. However as data is more than doubling every two years and as the top-tier market will get saturated, I am sure we’ll start seing offerings with more reasonable price-tags over the next few years.

* You may have notices that Microsoft is missing from this list – MS is also making moves in this field with Parallel Data Warehouse and Dryad . Unlike the other’s they do not have hardware capabilities as well as some other challenges. Maybe I’ll dedicate a post to them in the future.

Related posts:

Killing the HBase zombie table Random thoughts on big data Big data isn’t – well, almost