Data Warehouse Appliances – Should you consider one?

17 April 2013 / iReady

Recent industry surveys have shown a steady increase in demand and uptake of data warehouse appliances. According to The State of Data Warehouse in 2012 by Gartner, 90% of respondents considered investing in a data warehouse appliance. 42% of the 340 respondents in SearchDataManagement.com’s 2011 online reader survey with a data warehouse in place or planned within the next 12 months, indicated they were using or planned to deploy a data warehouse appliance.

In order to understand the uptake of analytic appliances it is important to note the paradigm shift that has shaped the information management needs of organisations of late. Organisations are struggling to keep up with computational needs of processing ever-increasing volume of data. Traditional scaling up and out of servers is no longer enough to keep up with processing and analytical needs. Analytical appliances offer higher computational capabilities either through a parallel processing or in-memory architecture. Vendors such as IBM, Teradata have solutions based on massively parallel processing (MPP) architectures whilst Oracle and SAP have in-memory analytic appliances.

Secondly, “greenfield” organisations have realized that delivering business value quicker is crucial to a successful information management program. Data warehouse appliances rapidly reduce the time required to scope, procure, setup and optimise infrastructure required for data warehouses. However, increasingly vendors have now begun offering data warehouse appliance solutions for mid-market organisations and departmental customers who have previously found the likes of SAP HANA, Oracle’s Exalytics platforms outside their budget. Mid-market solutions such as IBM Smart Analytics, HP Parallel Data Warehouse and Dell Quickstart Data Warehouse appliance are examples of solutions available for customers on a budget.

Should you consider investing a data warehouse appliance?

Apart from key use-cases listed above, there are a few other points worth considering –

Ease of use and setup: DW appliances are preconfigured, optimised hardware solutions designed to work with vendor software. This helps eliminate days if not weeks that would traditionally need to be spent in order to estimate the hardware requirements, assess compatibility, patching and software updates. Vendors typically spend millions optimising the appliances in order to get the best performance out of them for their software.

Simplified support and troubleshooting: As the hardware and the required database and analytics software ship together, the support model for appliances is relatively simple to work with. There are fewer variables in the environment to go through in the event of troubleshooting. This in-turn helps organisations reduce cost in on-going support and costs. BAU teams can be smaller with little to no infrastructure training required for existing team members.

Recently I was part of a InfoReady team engaged to deliver a Proof of Concept (PoC) using IBM Smart Analytics 5710. Our client had an urgent business need to implement a data warehouse with little IT support. Within an hour of unpacking the appliance, we were ready to pull out source data from the source systems and load the Data Warehouse platform. A few days later the foundation of the BI capability was set up. The box shipped with leading business analytics software, IBM Cognos Business Intelligence. Eventually, we were able to deliver a greater set of reports than our client anticipated. We saved weeks of trying to understand the required hardware, setting up and configuration. This allowed us to focus on client’s immediate information needs.

To summarize, organisations are becoming increasingly aware of challenges presented through having to manage large datasets and transforming them. It is also becoming clear that rapid delivery of business value is key to any information management program. Data warehouse appliances are now becoming more accessible, even for mid-market customers and must be strongly evaluated for your next project.

Previous  > >  Next