Blogs

Appliances are purpose-built

Sr. Technical Marketing Analyst, IBM

I’ve been working through my golden rules of appliances in my previous posts, and evaluating various vendor’s appliance claims as we go.

  • Appliances are Plug and Play
  • Appliances are purpose-built
  • Appliances are easy to use

Today, I’ll take a closer look at the second of these. Let's take a look at what this really means. By "purpose built," I mean that the solution was designed specifically for the task at hand. It is not a general solution that has been repurposed with new messaging in the hopes that it will work.

Many industries are guilty of this, but to drive the point home (no pun intended), let's take an easy example from the auto industry. The market trend over the past 5-10 years has been towards SUVs; big four-wheel drive vehicles that can handle off-roading and rugged terrain very easily. The car manufacturers realized fairly early on that very few of these "SUVs" ever went off-road, and that they could save a lot of money by building them on a standard car chassis. But if you were to take one of these vehicles on actual off-road terrain, it would be woefully inadequate to the task.

Appliances are all about using the right tool for the task. If that task is analytics for big data, it means that the solution is engineered from the ground up to do that task, and excel at it.

Of the solutions discussed here, one clearly stands out. Only Oracle was not purpose built for data warehousing and analytic workloads with its core roots always stemming from its OLTP strengths.

We see that clearly in the marketing spin from Oracle as it has evolved over the years. When Exadata v1 was released on HP hardware, Oracle claimed it was the best solution for Oracle Data Warehouses. Nobody really bought it, and when Oracle released v2 after the Sun acquisition, there was new messaging which went something like "Exadata is the fastest Oracle database for OLTP workloads. And by the way, it can also run data warehousing workloads as well."  Very different.

For those still keeping score:

Vendor

Plug and Play

Purpose Built

Easy to use

Score

IBM Netezza

X

X

 

2

Teradata

X

X

 

2

Oracle Exadata

X

 

 

1

ParAccel

 

X

 

1

EMC Greenplum

 

X

 

1

HP Vertica

 

X

 

1

Exadata v2 largely amounts to an attempt to shore up a 20th century OLTP database with some new technology (Infiniband, lots of RAM, semi-intelligent storage nodes) and the ability to bolt on some analytics components (Exalytics, Cloudera/Hadoop, etc.). Yes it is appliance-like, but all of the underlying complexity that goes into administering Oracle in general (indexes, partitioning, etc.), and Oracle RAC specifically, is there in Exadata and the general Exa* family of products. Exadata was not built from the ground up for analytics and data warehousing, and this is even more evident in the big splash Oracle recently made with their Big Data Appliance. Most of the players in this space have been doing Big Data for years and don't feel the need to shout about it.

The other players on our score card are undeniably built from the ground up for data warehousing and analytic type workloads, so get a check mark in the "purpose built" category. However, most of them do little to mask the complexity of running an MPP shared nothing big data cluster, which I will discuss in my next blog entry - appliances are easy to use.


For more information: