A colocation data center can provide IT facilities equipment, rack space, connectivity, and environmental support facilities, including power, cooling, physical security, and physical storage. Data center colocation service providers sell (lease) to both the wholesale and retail markets, and these services are typically priced on the rack displacement and the power presentation, in amps or kilowatts per hour.
The majority of colocation consumers choose a service provider based on a number of key attributes including conjunction with a high-speed communication hub, consumer to data center proximity, data center to data center proximity for bandwidth capability, data center tier and service accreditation (for example List X), regional locations or global presence, rather than just cost.
Unless the data center service providers take the opportunity to realign their services it will continue to drive down the original popularity and value of the colocation service, whilst other hosting-type services prevail, which will probably result in a reduction of customer retention and the potential new business.
WHAT ARE THE CHALLENGES?
The cost of on-premise data center environments increase based on energy, cooling, power management (UPS), fire protection, continued refurbishment, physical security, staffing, insurance, and in some instances, water management. The cost of alternative service provisions is reducing, however, only 1% of production platforms and applications have been transitioned to a public cloud infrastructure. Data center power consumption is 1.5% of the worldwide electricity output, costing $27bn (£17bn); and creates 210 million metric tonnes of carbon dioxide. Note: on-premise data center economics works for very small or very large organizations.
Additionally, the typical colocation service models do not provide flexibility for expansion, reduction, or the capability to flex-out for rack displacement or power consumption (during predicted or unpredicted business peaks) because the typical service contracts are positioned as property leases with fixed term, fixed price, and for some, fixed capability. A portion of this is driven from power presentation to large data centers, prohibiting scalability, growth, and capability.
Inaccurate assessments and customer guestimates of their power requirements force customers to predict and over-procure to the maximum capacity, including a risk percentage to ensure service availability. If future planning is not accurate the complexity of changing contracts, logistics, and connecting two isolated areas can impact the time to provision, creating risk and maybe increasing costs.
The colocation services market has suffered in favor of emerging and established hosting services, however, many organizations still favor colocation data center services as their requirements align to this service model, including the following requirements:
• Maintaining full control of applications and infrastructure in a trusted environment
• Data assurance and asset protection
• Regulatory compliance for data location and processing
• Unsupported cloud platforms
• Application functionality, complexity, and readiness
• Security concerns
• Desire not to convert premium cost office space to data center facilities
• Business strategy and policy limitations for business units, partners, and customers
• Cloud service model inhibitors such as application complexity, resistance to change, economics, and security
WHAT IS THE POTENTIAL ROADMAP TO CHANGE?
The economy has changed the way IT is procured and data center service providers should take the opportunity to differentiate themselves from their competition, providing more compelling reasons to use their services.
With the knowledge that customers still have requirements aligned to external data center services and choose to use colocation services, how can service providers mature their services and evolve the colocation market?
Service providers could innovate with the following:
1. The data center is simply a component of the IT service and should be procured, consumed, and treated as such, by both the service providers and customer communities, moving towards an IT-as-a-Service consumption model
2. Transparency— of consumer power usage, data center power consumption and availability, power changes, maintenance, and administration fees — via a customer (retail) or partner (wholesale) portal
3. Flexible contract periods: allowing IT service aligned contracts, examples include:
• Fully flexible; a rolling three month contract with one month renewal termination points
• Short fixed term with a rolling renewal; an initial six months and ongoing three month rolling renewals
• Medium term fixed, three to five years
• Long-term fixed, five years plus
4. Simple billing models: using standard fixed and Pay-as-you-Power (power consumption) utility models. Examples include:
• A fixed (guaranteed) power contract: catering for a predicted and well-defined power requirement where the customer pays a fixed price, irrespective of actual consumption
• An elastic (PAYG) power usage contract: catering for unknown and unpredictable power requirement where the customer pays for only the power consumed with no fixed or minimum amounts
• A committed power usage contract: where customers pay a combination of a fixed rate for a percentage of their perceived power consumption and an elastic rate for their peak power consumption
• A stepping power usage contract: using a rolling three-month baseline assessment (actively calculated and set each month) to determine an ongoing average power usage where small peaks and troughs are rolled in to the next two months average
• A minimal power usage contract: catering for standby, burst (into a colocation data center), and disaster recovery power consumption with the following additional capabilities within the billing models:
• A burst power capability: providing additional power consumption, using a fair-use policy where contention above the guaranteed rate occurs
• A space rental capability: allowing customers to acquire and guarantee adjacent rack space for future expansion
• A billing transition capability: enabling customers to move sections or contract between billing models but within the same contract
5. Intelligent power distribution units (PDUs) can be used to determine what outlets have provided how much power, with usage recording, predictability and peak/committed notifications, while using software agents to baseline and predict changes in power consumption and potentially assign power usage to internal customer cost centers
6. Professional services: for customer on-boarding assessments (space, power, and billing requirements) and readiness activities, improving the requirements accuracy, customer experience, transition governance, and increased time-to-consume
7. Monitor and predict energy price changes: notifying customers of projected power costs based on their historical consumption
WHAT ARE THE ADDITIONAL USE CASES?
With the data center services improvement, the following additional colocation use cases become feasible:
• Rapidly expanding businesses or IT services, including start-up organizations
• Multiple development, test, and pre-production environments, where enterprises have multiple and staggered development cycles
• Disaster recovery environments, with the majority of recovery platforms powered off or low consumption, during normal operation
• Storage only environments to support backup, archiving, data protection, and big data initiatives
• Complex application and infrastructure environments
• Flex-out resource environments from an on-premise data center to increase capacity during predicted and unpredicted peaks on IT infrastructure
• Heterogeneous cloud and colocation environments to support business IT strategy/model and migration or transformation scenarios
• Return the use of premium cost office space to the business
WHAT ARE THE BENEFITS?
Many benefits could be realized by an evolution of the colocation service offerings, for both the service provider and consumer communities, at a financial, development, and innovation level.
With the capability to use less power and benefit from consumption only billing, this will encourage consumers to assess usage and become power efficient, by introducing power efficient technologies and effective power management policies, within a fit for purpose data center environment with a lower PUE. This could be achieved by manually or automatically powering down unused platforms. In turn this behavior should see a steady decrease in the overall cost of IT services to the business.
The ease of steady consumption and flexible consumption billing models will remove the risks of performing rapid migrations and associated high-costs when transforming an infrastructure.
This re-invigoration would naturally lead to customers consuming more cost effective data center colocation services by allowing service providers to fill their data center space, improve annuity revenue, and drive continuous innovation, improvement, and efficiencies. There is potential to introducing additional revenue generation through professional services, managed services, and service readiness, and customer retention based on quality services from the data center operator.
The level of competition will hopefully lead to global data center efficiency and innovation standards and certifications, which are used by customers when selecting potential service providers.
CONCLUSION
The widespread underutilization of data center space and power consumption will decrease and become consumed in the controlled and secure manner required by the customers’ business while driving efficiency and data center standards and stimulating the data center market.
The inevitable outcome for consumers will include a lower cost of ownership. Among all the benefits of the points of evolution, the most significant is that it will drive good behaviors for green and clean data center hosting, hopefully leading to a reduction in power waste, IT service cost, and IT’s impact on a businesses’ carbon footprint.
