Teradata & Cloud - Last Roll Of The Dice?

Teradata & Cloud

'Happy New Year' To You Too!

The Christmas & New Year break (not 'holidays', ugh!) is a time for a bit of rest & relaxation away from work, eating & drinking too much, pretending to like spending time with your relatives, swapping presents and generally winding down.

It's also good thinking time...In addition to 'messing about with databases', especially Teradata, for the best part of 30 years (yikes!), yours truly also has more than a passing interest in personal finance and investing...which got me thinking about Teradata's share (stock) price performance, obviously.

Teradata's Share Price

Teradata share price 2007-2012
Teradata 2012-2016

After splitting off from NCR, Teradata floated via an IPO on the NYSE and closed at just under $28 on the first day of trading on 1st October 2007, or 1071001 for the TD geeks. Over the first 5 years Teradata's shares rose a whopping 200% (source Google Finance):Since the heady days of Q3 2012, when Teradata peaked at around $80, it has been a rather different story:

According to Google Finance, as of 6th Jan 2016, Teradata's share price has declined 68% since the peak in Q3 2012. There have also been declines of 44% in the last year, 31% in the last 6 months, 17% in the last 3 months, and almost 10% in the last week alone.Update 14th Jan 2016 - since I started writing this article last week Teradata's price has declined by a further 2% or so to under $23.

Those that bought via the IPO in 2007 are now looking at paper losses. This is hardly a stellar performance from the world's biggest pure-play analytics company.

Why Has Teradata's Share Price Fallen?

There's only one reason that share prices fall: more sellers than buyers. It's that simple.What motivates sellers to sell, and buyers to buy, is an entirely different question and one that we can only speculate on.Efficient market theory would have us believe that all known information about the future potential of a company is reflected in the share price. This leads to the conclusion that 'the market', whoever that might be, hasn't held out much hope for Teradata's fortunes for over 3 years, since the 2012 peak.

What Has Happened to Teradata Since 2012?

The short answer, in a single word... Hadoop.Teradata's historic value proposition was quite simple: massively parallel processing (MPP). The ability to deal with data volumes that others struggled with through a 'scale out' MPP architecture set Teradata apart from the likes of Oracle, IBM, Microsoft etc who relied on the traditional 'scale up' SMP architecture.

For the best part of 25 years from the late 1980's to the early 2000's Teradata was alone in the enterprise market with its eponymous MPP database offering. Then along came the likes of Netezza and Greenplum with competing MPP database offerings. Netezza at least caused Teradata to sit up and take notice, before being bought by IBM. Greenplum was aquired by EMC and then spun-off as part of Pivotal, but has hardly had any impact on Teradata's fortunes.

Teradata's response to Netezza was the launch of the more affordable range of Teradata appliances to complement the existing Teradata enterprise models.

However, the new MPP database players collectively turned out to be minor irritants compared to the impact of Hadoop on Teradata. Our first hand experience is that, rightly or wrongly, many enterprises see Hadoop as an alternative to Teradata's MPP database. The fact that so much effort is being spent to run SQL on Hadoop, and in doing so turn Hadoop into what is basically an MPP database, is deeply ironic.

Why not go with an MPP database system in the first place, given that it will cover 99% of use cases for 99% of companies? But that's a whole other story...Teradata's response thus far to Hadoop has been less convincing. The strategy seems to be to promote the Teradata Unified Data Architecture (UDA). The UDA consists of 3 elements: the traditional Teradata data warehouse (production), the Teradata Aster appliance (exploration) and a Hadoop appliance ('Big Data'). Confused?!?!?!

Teradata on Amazon AWS

The availability of Teradata via the public cloud, and specifically AWS, was announced in late 2015 for availability in Q1 2016. Teradata's more recent thoughts on Teradata in the cloud are neatly covered in a BeyeNETWORK article.

As Teradata's Brian Wood points out, Teradata in the cloud means "Teradata Database is opened to a much broader range of potential customer".

While this is clearly true, key questions remain: how much interest will there be for Teradata on AWS? Will multi-node Teradata MPP systems work satisfactorily on AWS? Will Teradata on AWS be profitable for Teradata?

One thing is for sure, Teradata needs to do something to arrest the precipitous share price decline. Maybe the launch of Teradata on AWS, and other public clouds such as Azure, is that something? Only time will tell, and that time will be upon us soon.

As a Teradata professional services provider, the folks here at VLDB are crossing our fingers in the hope that Teradata on AWS is both affordable and can run at scale beyond a single node. More Teradata users is good news for us!

These are interesting times indeed for Teradata.