Many corporations and other large operations that previously would only consider owning and operating their own data centers are now considering moving at least parts of their enterprise into collocation sites. This major paradigm shift in strategy raises some very interesting new challenges. Entities that previously had direct control over every aspect of their critical operations must find new strategies for ensuring continuous operations for their corporate endusers.
Traditionally, most large enterprises constructed and operated purpose-built data center facilities based on specific needs and requirements. Site selection was influenced by proximity to other corporate locations, cost of utilities, broadband connectivity, and a host of other parameters. The design and engineering solutions employed reflected internal needs, preferences, and technologies that existing staff and management were most comfortable with. Decisions regarding the level of mechanical and electrical redundancies, capabilities to sustain critical operations during utility outages, and monitoring and control systems were often a balancing act between the best use of capital funds and providing the reliability and availability required by specific IT needs.
Just as important as the decisions regarding the design and construction of the physical facility were the decisions regarding how to operate and maintain the data center. CIOs needed to decide how to staff the facility including if 7x24 coverage would be required, how many personnel to put on each shift, and their respective competencies. Decisions regarding which tasks and activities to perform in-house vs. outsourcing would drive the development of site-specific operating protocols, O&M programs, and training requirements.
The overall success or failure of the mission hinged on how well these decisions were implemented. This meant corporations and other entities that are otherwise completely unrelated to the IT industry had to also be experts in the design, construction, operation, maintenance, and overall management of data centers. Not only did banks have to have financial experts but they also had to have experts in emergency generators and critical power distribution. Department stores and retailers had to have experts in the operation and maintenance of complex central chiller plants. Basically every mission critical operation had to be expert in both the “mission” and in data centers and critical facilities.
The obvious advantage was the ability to control their destiny. Corporations that own their data centers have direct control on how they are operated and maintained. The downside of course is that it takes a great deal of resources and effort, including advance planning and both capital investment and recurring operating expense.
Moving an enterprise to a collocation site has many advantages. Many collocation companies build new or expand existing data centers on speculation that demand for their services will continue to grow. This “build it and they will come” strategy results in “white space” being available to prospective endusers at any given time. Rather than wait a year or more to construct and commission new space, move in can start as soon as an appropriate site is identified and leasing contracts and service-level agreements are signed. The corporations and their enterprises can focus on their core business and leave all that facilities management work to the collocation’s staff to worry about. Or can they?
Is the impact of an IT outage any less when it occurs in a collocation site that is an owned site? Is the loss of reputation any less? The loss of revenue? So the down side to moving an enterprise to a collocation facility is the loss of direct control over the quality of design, construction, operation, and maintenance of the data center. The corporate need for reliability and availability are the same but the means to achieve them are different and so this drives a new set of best practices based on oversight and continuous due diligence.
Not all collocation sites are the same. Collocation companies offer a range of capabilities and services. The sites vary in size, technology, staffing, and most every other characteristic. One aspect that they pretty much all have in common though is that their corporate bottom line depends on maximizing the rent collected while minimizing expenses. This can present some fundamental conflicts between their goals and objectives vs. their client’s expectations for sustained operations. This means that corporations that migrate to a collocation strategy need to also migrate their critical facility focus from an execution perspective to an audit and oversight perspective.
Before an audit and oversight program can be put in place it is first necessary to establish specific performance requirements and criteria as well as consequences for failure to perform. At one extreme, the service-level agreement between the collocation company and the enterprise could basically require 100% uptime of all power and cooling for some value of watts per square foot of load in the leased space and define the consequences for failure. The problem here is that rarely will the consequences for failure ever cover the actual costs incurred by the enterprise.
A better approach would be to include in the leasing arrangement the specific facilities management and O&M performance requirements that meet the enterprise expectations. This would include most of the same requirements that would be applied to the facilities management department of an owned data center. These requirements should cover O&M staffing, alarm response and escalation, equipment inspection and maintenance, use of methods-of-procedures (MOPS), and standard operating procedures (SOPs). Performance criteria should be defined including formal key performance indicators (KPIs) such as compliance with ASHRAE TC9.9 Thermal Guidelines, IEEE, and NETA maintenance standards, and managing loads below safe thresholds to preclude cascading power failures, etc.
The key to success when migrating to a collocation strategy is to “inspect what you expect.” This starts with doing full and comprehensive due diligence on a prospective site before negotiating a lease. Consider hiring objective critical facility experts to perform a site reliability assessment that evaluates the design and construction of the facility. Almost every data center built in the last few years has been formally commissioned, so include a thorough review of the final commissioning report. Some collocation sites reduced costs by purchasing lower quality materials and equipment and doing just enough quality control to claim the site was “commissioned “ (or what is referred to as “commissioning light”). Once the contracts are signed and the move is completed it is almost impossible to reverse course.
The due diligence and quality control should continue for the entire occupancy period. The lease agreement should stipulate that ongoing performance audits and inspections be allowed including on a random and short notice basis, if possible. A formal quality control program should be established that includes review of incident reports, completed maintenance activities, staff training, etc.
Most collocation companies understand that their success is hinged on providing high reliability, flexibility, and responsiveness to their client’s needs. They may not like the additional oversight, but they shouldn’t fear it either unless, of course, they have something to hide.