Numerous telcos, like Verizon, CenturyLink and Tata, have publicly said they are evaluating the feasibility of selling off data center assets. This seems to have created a flurry of hasty conclusions that ‘the data center is dead’.
We saw this assertion previously beginning in 2012 around talk of the demise of the data center due to the rise of cloud computing. But as we know now, the cloud simply changes where the applications are running. It all goes to a data center somewhere. And it is clear in 2016 that the need for strong data center operations is as critical as ever, perhaps even more so.
The decision for any organization to sell its data center assets belongs to the Chief Financial Officer. This is when getting an asset ‘off the books’ becomes a catch-all for a variety of motivations, and involves depreciation cycles, cash flow, capital reserves, and assuring shareholders that an organization is only ‘carrying’ assets that are core to its business. Be assured that these specialists are not selling data centers because they are no longer valuable to their business.
Regardless of the final commercial transaction, such as a sale-leaseback, the data center operations may be unaffected. A real estate transaction does not necessarily change the configuration of the equipment. And, preliminary insight into some of the telcos considering to sell their data centers is that the staff will not likely change. I’d also say it’s perilous to assume that a change to the name on the building title changes everyone’s IT performance requirement.
+ ALSO ON NETWORK WORLD 10 more of the world’s coolest data centers +
Enterprise IT typically demands more of their outsourcers than from their own team. An enterprise IT organization may readily admit the shortcoming of its team, but would certainly not select a service provider with holes in their capabilities. A change to an outsourced model, even by title alone, will bring increased scrutiny on the effectiveness and efficacy of data center operations programs. In the last few years, other parties have entered this space with third-party certifications developed specifically for data center operations. (Disclosure: I am employed by Uptime Institute, a company with an offering in this regard.)
Agnostic to real estate transactions, the choice to outsource operations to a specialized management company (such as CBRE/Norland, JLL, Grubb & Ellis, & etc.) is more prominently seen, and currently being initiated by some of the most respected and well-resourced DIY enterprise IT groups. (Or perhaps I should say OIY: Operate It Yourself enterprise IT groups.)
You can dangerously assume that all service providers have your best interests, and superlative performance, as their business model. Or that in this configuration that there is no need for oversight. This short-sighted behavior is not uncommon, and ironic in organizations that have once- or twice-a-year due diligence assessments of their data center infrastructure, which is likely static.
Yet, they do not have the same for the operations, which is highly dynamic and represents a 3x risk exposure than the infrastructure itself. Data center equipment does not change year-over-year. You may have end-of-life replacement, but that occurs approximately every 10 years. In the event of a capacity expansion, this will lead to more equipment being installed, but not a fundamental change to design philosophy. However, staff turnover may be multiple times in a year, and an operating philosophy can be strongly dictated by the person leading it.
The important reality is that data center management is a business, and must be both practical and profitable. To remain viable and competitive, your management company must look for ways to do more with less. Shared resourcing is more common than you may think, meaning your 7 x 24 might be the same person at one or more other sites. Do you know your 7 x 24 expectation is actually a 1-hour response time SLA because it is cheaper for your operations company? How do you protect your operations program (whether outsourced or OIY) against such cost shavings measures?
+ MORE ON DATA CENTERS: The world's coolest, greenest data centers +
Here’s a story about a recent assessment completed for a global systems Integrator who had recently switched management companies. As the prior company departed, they took the entire library of processes and procedures with them. They believed the library was part of their service delivery and therefore transportable. The incoming company was in the midst of figuring out how to rebuild the library when the site was dropped due to an error in manual sequence of operations. I find this example useful because the systems integrator is in the IT business themselves and they struggled with the result. Part of this is attributable to selection of lowest cost provider, which drove the change in management companies.
Alternately, your outsourced operations company may be over-charging by not seeing or putting into action opportunities for more efficient staffing. If you are an OIY, these are just as applicable. We all need to be mindful of overall management execution, over-charging, under-service, etc., even within our own teams.
Here are some common indicators and questions to ask to better understand if your operations team is delivering value or underserving your IT organization:
- Does your staff sit around waiting for the next work order or maintenance activity?
Staff underutilization may be addressed by assigning staff to perform rounds and appropriate maintenance activities, thereby reducing headcount or vendor costs.
- Have you analyzed your maintenance frequencies?
The possibility of revising some frequencies, as consistent with best practices and manufacturer tolerances, can reduce maintenance costs.
- Do you use tracking tools for work orders?
Labor hours can be reduced by ensuring maintenance is not started until all tools and materials are on hand.
- Do you check work orders against warranties?
OPEX funds could be wasted by repairing equipment under warranty.
- Do you track incidents and trend equipment performance?
Attempting repairs again and again may be wasteful.
- Do you compel your vendors to hold parts or do you accept the costs of inventory?
These costs could potentially be transferred via improved vendor contracts.
- Your data center set points may be excessively provisioning your environment, and the squandered expense is incurred continuously.
One recent example of an operations issue is a recent cloud outage that made both industry and general news. The outage was initially reported as a failure of the service provider. However, upon root cause investigation, it was revealed that the Internet exchange through which multiple countries accessed the cloud provider had failed. The failure was the result of a back-up generator failing to take the load.
From an executive standpoint, some of these solutions may seem tactical or too obscure to be actionable or even hold their attention. But, executives need to be asking for, and receiving, a shorthand to be assured that the data center operations are robust regardless of whether they are outsourced or OIY. There are no insignificant activities when you are dealing with an imperative need for high availability.
With the growing importance of data, 24x7 support has to be provided as our “always on” society demands it. Even for those who wish to distance themselves from the data center conceptually or operationally, there remains a data center somewhere. And whether the data center is company owned or sold to a service organization, efficient management and collaboration to keep everything running smoothly is imperative to both parties.
This article is published as part of the IDG Contributor Network. Want to Join?