ENTERPRISE MANAGEMENT ASSOCIATES®
by Andi Mann, Research Director, EMA
Despite its resurgent popularity, Data Center Automation (DCA) is not by any means a new technology area. As far back as the late 1980s (and even earlier), technologies were developed and implemented to automate various aspects of mainframe computing. In the early 1990s DCA started to become more popular with the rise of proprietary products from DCA pioneers such as Candle Corporation (now part of IBM), Goal Systems, Legent Corporation, UCCEL (all now part of CA), 4th Dimension Software, Boole and Babbage (both now part of BMC Software), and others. Over time, DCA also accommodated UNIX, AS/400, VMS, Tandem, etc. Later still, it adapted to automate the management of Windows, and most recently, Linux servers.
With this history in mind, EMA defines DCA as “a set of IT management disciplines aimed at automating system and application management activities within medium- or large-scale centralized computing environments.”
This broad definition contrasts with the sometimes-limited views held by some technologists. However, EMA analysts – who are all enterprise IT practitioners with extensive personal experience working in Data Centers of various sizes – understand the full history of DCA, and realize that it spans as many different disciplines as it takes to automate the management of a Data Center.
While not a definitive list, EMA considers at least 15 different disciplines to be core to DCA. These include: provisioning, patch distribution, change and configuration management, virtual machine management, asset and inventory management, compliance auditing and reporting, remote control/out-of-band, event management/console automation, automated backup and recovery, database automation/optimization, capacity management, job scheduling/workload automation, performance and availability monitoring, IT process automation, and charge-back automation.
With this broad focus, EMA has just completed the most wide-ranging primary research undertaken to date into the world of DCA. Using an extensive survey, one-on-one interviews, and individual case studies, we analyzed the primary use cases, functional benefits, business drivers, outcomes, penetration levels, importance, cost benefits, satisfaction, response time and availability impacts, and barriers to adoption for a wide number of DCA disciplines, technologies, and solutions.
This research – published as a full EMA Research Report titled “Data Center Automation: Delivering Fast, Efficient, and Reliable IT Services” – analyses over 300 separate data points collected from over 200 respondents.
Some of the more interesting highlights include:
- DCA lives up to, and in most cases exceeds, all expectations. Customer satisfaction with DCA solutions is high, both for individual disciplines, and across the entire catalogue of technologies. Over three quarters of respondents reported that DCA provides real and measurable profitability and cost benefits. While the upfront costs of DCA solutions are a concern, the other significant barriers to deploying or expanding DCA implementations are not product-related, but instead are internal and human issues like politics, lack of resources, and lack of processes.
- Data Centers are primarily deploying DCA for strategic reasons – not just to save money. Sure, the cost savings are nothing short of amazing; according to the research, DCA technologies effectively halve the staffing requirements in the Data Center, saving the average business almost three quarters of a million dollars in staffing costs alone. However, the primary drivers for DCA tended strongly to be strategic, focused on priorities like getting away from reactive “firefighting” so administrators can use their skills on project work instead, improving response times to meet more stringent SLAs and improve business service, and providing better data security in order to reduce risk.
- Alternative form factors for DCA solutions are on the rise – but are not dominant yet. As long as available solutions have equivalent functionality, most decision makers still prefer a suite-based software solution, with multiple disciplines all in a single package. However, a surprising amount – almost a quarter of all respondents – expressed a preference for open source software. A substantial percentage actually preferred a hardware-based appliance (you can check out the EMA Advisory Note, "The Pros and Cons of Systems Management Appliances" for more background on the appliance model). Data Centers may be focused on strategic outcomes, but they are also hoping to deploy DCA solutions in less expensive form factors.
- The myth of the “lights-out Data Center” is finally and officially “busted.” Even with sophisticated DCA technology, it is still far from a reality for most Data Centers. While a handful of Data Centers may have achieved this lofty goal, according to our survey it is the least important of all possible drivers for DCA deployment, and is the least effective of all possible outcomes of DCA. In fact, almost none of the respondents even want unattended automation, even with an ideal DCA toolset – let alone with current tools.
Of course, at well over 20 pages, the complete research report has a great deal more data and analyses, and goes into much more interesting detail than I can possibly put in this newsletter. It provides in-depth analyses of the use cases, importance, outcomes, and satisfaction ratings for each of EMA’s 15 core DCA disciplines; highlights the key vendors in each area; looks at the impact of DCA technologies individually and collectively on key metrics like Mean Time To Repair (MTTR) and Availability; analyses the potential cost benefits and staffing impacts of DCA; and much more. It also contains several case studies of DCA in action in the real world, and wraps up with EMA’s perspective on the broad picture of the benefits, futures, and concerns regarding DCA technologies.
I do hope you will get a copy of the full report. At the very least, you should certainly catch up with the free replay of my DCA Webinar entitled "Data Center Automation: A How to Guide to an Effective Getting-Started Strategy." Of course, if you want to consult in person with me or with any of EMA’s analysts on this or any other topic, or even just share your own experiences of DCA, please feel free to get in touch. You will find the DCA report,the DCA Webinar, and all our contact information at www.enterprisemanagement.com.