September 08, 2010
Perhaps a better title would be, “yet another reason why mission critical applications are staying put” since this is so frequently a topic of discussion. However, to break from the string of latency, performance and development challenges for a moment, we are left with the more nebulous issue of IT culture; namely, the clearly defined tensions between virtualization advocates and the stalwart owners of core, revenue-critical applications.
Application performance management firm AppDynamics released a survey that revealed “a significant divide in the pace of adoption of virtualization for non-critical systems versus mission-critical systems.” More specifically, the survey results suggest that only 14% of the approximately 3500 large-scale enterprise leaders surveyed virtualized any of their essential operations, in part due to worries about application performance degradation but also in part due to some larger cultural issues that are the core of such decisions—a point that we’ll explore in greater depth momentarily.
The survey asked another layer of questions to that slim 14% of enterprise IT leaders who stated they had already virtualized mission critical applications by inquiring about the nature of those applications. For nearly all of the respondents in this small minority, the mission-critical applications that were actually virtualized were overwhelmingly employee-facing versus customer-facing, which still indicates at least some degree of trepidation about the reliability of virtualizing anything that led to profitability. As Roop stated, “if there is a performance degradation that comes with employee-facing applications, the company is willing to bear it and figure the employees can do so the same, especially if there are clear cost savings associated with virtualization.”
Without a doubt, virtualization, while heralded as the best technological development for cost savings since barcodes on coupons, does strike fear into the hearts of many IT stakeholders. The virtualization administrators need to deliver on these cost promises, the application owners need to be able to feel that they will be able to deliver the same level of performance and uptime as usual—if not better, and the rest of the organization needs to be able to keep pace with the change. However, this change is clearly very slow in coming—but is it all based on security, multitenancy, and the host of other dirty words in the cloud ecosystem or is it really just about culture and a lack of communication?
The Cultural Factor in Virtualizing the Mission Critical
It is difficult to find extensive case studies of large-scale enterprises that have virtualized any significant part of their mission-critical operations, in part simply because if they have they might not be particularly excited to share these details with their competitors, but in larger part, because there are several barriers preventing any wholesale move to a virtualized infrastructure, not the least of which is general IT culture itself and the performance standards it sets within its firmly entrenched camps.
For every virtualization administrator there’s an application owner that must be convinced that his or her application would thrive (or even function) off the physical hardware. And since the two parties have little in common outside of the fact that they might be working for the same CEO, there’s naturally a bit of tension present, due in part to a lack of communication about what is at stake for both parties and how they might share tools and measurement standards to thoroughly evaluate the possibility of application virtualization.
According to Steve Roop of AppDynamics who spoke with us following the release of the survey, this struggle in IT departments is based far more on tools, measurements and accountability rather than a fundamental divide. Since a virtualization admin’s superiors rate their underlings on the level of virtualization achieved and an application owner’s superiors look for constant uptime and SLA adherence to high performing core applications, it’s no wonder why these two camps have become impenetrable silos—both retain (and perhaps guard) their specialized knowledge and have no understanding of the measurements and standards by which the other functions.
On the surface, all that might be at issue here is a simple lack of communication on both sides, but if that issue has already been recognized and there’s still no progress, Roop suggest that perhaps a more fine-tuned approach is necessary to help the virtualization advocates convince the application owners that they are onto something useful (and even beneficial) that fits with their requirements while conversely, the application owners open their silos to help virtualization administrators understand what their unique concerns are.
Consider the following statement from the survey: “52 percent of companies noted that application owners have blocked the project to virtualize a Tier 1 application. At the same time, 49 percent worry that the applications aren’t designed to support virtualization and 45 percent cited concerns about performance degradation once virtualized. “ While these sound like respectable technical reasons for concern on the part of the application owners who have very stringent demands for performance and uptime (among a host of other deliverables to constantly contend with) Steve Roop suggests that this is all still rooted in culture—in an unwillingness to open the silos and explore new ideas.
The main takeaway from our survey really is in this “people” issue said Roop. “53% said the number one obstacle was that application owners were blocking the virtualization projects. Now, 45% said that the underlying coenrerns were performance degredataion and that’s a reason for it certainly, but a lot of this is rooted in the cultural divide and so performance and those other worries have almost become secondary. “
In Roop’s view, “Both of these issues need to get solved at the same time for application owners to even say “yes, I’m willing to just give this a test.” In the past, application owners had the power to put their hand up and say, “only over my dead body will my mission-critical application be virtualized. I want to keep it on dedicated hardware and control everything.” What needs to happen then, at least in this line of thinking is that, as Roop put it, “the virtualiaation folks have to walk a mile a shoes of the application owners to understand their tooling and systems and how they’re measured so they might be able to show them that performance might even get better and if using the right tools, give some degree of confidence that it might work or even improve performance.”
The survey results were released at VMworld last week where Roop held his conversation with us. He noted that despite listening to a litany of liturgies aimed at virtualization administrators and how they can better bring application owners into their fold, there was no real discussion about the “silo” issue and how it might be at the root of the virtualization adoption problem. He claimed to have had a number of off the cuff discussions about this very issue in general conversation, but overall the conference seemed to avoid this gnarly topic.
In honor of the issue at hand, AppDynamics allowed Darth Vader to speak on the behalf of the application owner, harkening back to the “you’ll virtualize my application over my dead body.”
And that's Steve Roop, by the way. Vader sort of makes him look smaller than he probaby is.
But It Isn’t So Simple
AppDynamics, given its line of business, has clear interest in the issue, of course but still, it remains to be seen how much of this is conjecture and how much of it is based on the sense throughout an entire IT department that such core applications are simply not ready for virtualization on a performance and technical level—leaving behind any of the infighting among people.
There are plenty of examples of application degradation, as many are already aware. However, for cost savings and IT consolidation to become a reality, there has to be some clear incentive for application owners to at least take a look at how their applications might perform in such an environment. The question is, could the mere act of experimentation for the sake of possible innovation be enough of a driver alone?
May 23, 2013 |
The study of climate change is one of those scientific problems where it is almost essential to model the entire Earth to attain accurate results and make worthwhile predictions. In an attempt to make climate science more accessible to smaller research facilities, NASA introduced what they call ‘Climate in a Box,’ a system they note acts as a desktop supercomputer.
May 16, 2013 |
When it comes to cloud, long distances mean unacceptably high latencies. Researchers from the University of Bonn in Germany examined those latency issues of doing CFD modeling in the cloud by utilizing a common CFD and its utilization in HPC instance types including both CPU and GPU cores of Amazon EC2.
05/10/2013 | Cleversafe, Cray, DDN, NetApp, & Panasas | From Wall Street to Hollywood, drug discovery to homeland security, companies and organizations of all sizes and stripes are coming face to face with the challenges – and opportunities – afforded by Big Data. Before anyone can utilize these extraordinary data repositories, however, they must first harness and manage their data stores, and do so utilizing technologies that underscore affordability, security, and scalability.
04/02/2012 | AMD | Developers today are just beginning to explore the potential of heterogeneous computing, but the potential for this new paradigm is huge. This brief article reviews how the technology might impact a range of application development areas, including client experiences and cloud-based data management. As platforms like OpenCL continue to evolve, the benefits of heterogeneous computing will become even more accessible. Use this quick article to jump-start your own thinking on heterogeneous computing.