Last year, I discussed how excited I was by HP’s introduction of advanced telemetry in its servers and how having all kinds of information sent automatically to HP support could improve overall reliability and availability. More recently, I wrote an article about Nimble Storage’s expansion of its in-array telemetry, which, like HP’s solution, sends all kinds of data points back to Nimble HQ on a regular basis. According to Nimble, each arrays ships around 30 million data points per day to the company’s centralized data warehouse. From there, the possibilities are endless. Here are five ways that this kind of telemetry can be used to benefit the enterprise:
Have you ever had a situation where you’re picking up the phone to call someone and that person is already on the line because he called you first? Now, imagine this happening with an array support vendor. Even before a customer calls HP or Nimble, these companies have all the information an may have already opened a case. Other companies, such as NetApp, have been doing things like this for a long time as well. There are stories of customers receiving hard drives from NetApp and not knowing why. Upon questioning, they learn that a drive had failed and the array had notified NetApp, which sent a new drive. In a traditional support arrangement, the customer often has to run a series of scripts to generate a support file, which is then uploaded to the vendor before the vendor will take action toward resolution. This slows down the resolution process and is a general annoyance.
Over time, this level of support should become a baseline expectation for CIOs. It’s that important. The less time that it takes to go through the first line of support, the less downtime a customer experiences.
Root cause analysis
With so much data coming in, it might seem overwhelming. But, over time, companies that build these kinds of support information warehouses will be able to build enviable support databases that are based on real-world findings and that help customers identify and even avoid problems. With the right information coming in and the right people analyzing the information, root cause analysis can be performed much more quickly.
Not every software update provided by an array vendor is perfect, particularly as the vendor’s customers become more varied in their use cases and stray from baselines. Eventually scenarios arise that the vendor simply didn’t envision that could cause problems.
But, consider this: With every single array in the field sending back millions of data points, Nimble has particularly keen insight into every customer environment. So, once a root cause is determined for a particular problem, it’s likely that the company can quickly identify other customers who might be impacted by the problem and head off the problem.
Again, for CIOs, this level of service should be sought whenever possible and reasonable. It keeps internal staff from having to become storage sleuths and places more of the support burden on the vendor, which is where it should be.
What if analyses
Having the ability to be proactive in support is a powerful differentiator, but what if array vendors could take things one step further? With enough data flowing into the analytics environment, they may ultimately be able to predict the future. Organizations can work with their vendor partners to ask questions like “What will happen if I add VDI to my existing workload?” Because that vendor partner has deep understanding for the existing customer environment and also has deep understanding for VDI environments run by other customers, a reasonable impact analysis can be performed. Such analyses can greatly improve the chances of success of new rollouts. More importantly, this analysis can be accomplished without having to build what can be a costly and time-consuming proof-of-concept environment.
Action Item: Modern infrastructure components are becoming much more self-aware, thanks to the efforts of some vendors building complete telemetry into their hardware and software. At the same time, CIOs – as always – are being asked to do more with less. Wherever possible, CIOs need to shift workload from internal IT staff to vendors and maximize the support contracts they have in place. Rather than run proofs of concept and pay staff to do baseline troubleshooting, CIOs should insist on implementing solutions that can eliminate non-value add activities from their organizations. One way to do this is to implement solutions from vendors that have learned to harness customer data to improve support operations.