IT Focus Area: infrastructure operations
November 16, 2015
Find Untapped Opportunities in Your Data Center
Here’s an eye-opening statistic: Almost 30% of physical servers in data centers sit idle.
A report from Anthesis Group, Stanford University and TSO Logic found that 10 million global servers are “comatose.” They haven’t delivered information in at least six months but have cost companies more than $30 billion.
That’s a huge drain on your resources—both in terms of cost and the amount of time you spend managing equipment that you’re not using.
Although enterprises are moving their servers to colocation data centers to cut costs and gain efficiencies, paying for idle servers defeats the purpose.
You can optimize your data center in two key areas: the facility side and the IT side. Let’s explore your opportunities in both of these areas.
How to optimize your data center facility
When it comes to optimizing their data centers, many enterprises make big mistakes. Here are three keys to optimizing your facility:
1. Go vertical, not horizontal
Some companies think that expanding their data center means purchasing more floor space. However, you’ll be more efficient if you maximize your vertical cabinet utilization.
Most data centers only offer 4kW to 5kW of capacity in their cabinets. But with today’s high-density IT equipment, just one device can consume this much power. If your cabinets have low power densities, you’ll only use part of your vertical rack space before the cabinets run out of power.
Instead of expanding your floor space, aim to increase your cabinets’ power densities. In a higher-density data center, you can support the same amount of IT equipment with fewer cabinets and less floor space. You’ll also need less structured cabling, power distribution units (PDUs) and power whips. With cabinets and their infrastructure costing between $10,000 and $12,000 each, restacking multiple low-density cabinets into a single high-density cabinet can bring you huge savings. You’ll also have less equipment to manage.
Migrating to a new data center lets you start fresh and optimize your cabinets. However, you can also save capital expenditures (CAPEX) when you consolidate the cabinets in your current data center. For example, instead of expanding your data center, restack the IT infrastructure from several low-density cabinets into a few high-density cabinets. You’ll also improve your energy efficiencies and cut your power costs.
2. Optimize your structured cabling
Avoiding a giant mess of cables means doing things right from the start and then maintaining it as you go forward. Here’s how to optimize your low-voltage and high-voltage cabling:
Low-voltage telecom cabling
Many data centers have a rat’s nest of low-voltage cabling. You can eliminate this mess by cutting and running all your copper and fiber cables from the starting point to the end point. Comb all of your copper and fiber bundles so they don’t kink. This is important with fiber cables, as they won’t operate properly if they are kinked.
Mark each cable’s “to” and “from” end, so you know where everything is going. And don’t ball up your cables and toss them into cabinets. Make sure they are cut and run to length in each cabinet.
It’s crucial to properly document and manage your high-voltage cabling, as the safety of your data center’s staff may be at risk when these cables are not well managed.
The biggest mistake with high-voltage cabling occurs in the data center’s distribution panel. You can fix issues by documenting your A/B power infrastructure. Color code your cables, so you know which ones are for A feeds and B feeds. Putting arc-flash labels on each piece of equipment will keep your operations staff informed and increase their safety.
Your structured cabling design should be flexible enough to support new equipment, so you won’t need to add fiber or copper cables on demand. Adding cables every time you get a new device is pricey. You’ll also end up with a tangle of cables that is hard to manage.
Doing this extra cabling work upfront allows you to find and fix problems quickly. In the future, you won’t need to weed through cables to find the source of a problem.
3. Beware of overprovisioning
One of the biggest data center wastes starts in the design phase by overprovisioning cooling and power. Data center managers may plan for a certain load and then purchase the power and cooling equipment to support it. However, their actual IT environment may only need half of this load. Then, they’re stuck with extra equipment that they need to maintain.
Overprovisioning can also make your equipment operate less efficiently. For example, if you run an air conditioner at a lower load than what it was designed to handle, it will consume more energy than it needs to support the low load.
Another mistake easily made is relying on IT equipment’s nameplate ratings. A data center manager may look at the nameplate to see how much power a device is rated to consume and how much heat it exhausts. Then, they add all of their device ratings together and plan their power and cooling accordingly. However, IT equipment will most likely never run at 100 percent of its rated power. If you base your power and cooling requirements on nameplates, you’ll significantly overprovision your data center.
Instead of overprovisioning, start small and scale. For example, you can pre-provision pipe work for future equipment and then add the equipment as you need it.
You should also take the time to understand the business purpose of every device. You may have equipment in your data center that supports unused apps, which results in neither the equipment nor the apps serving a business purpose. If you tie each piece of equipment to its business purpose, you can eliminate waste.
How to optimize the IT in your data center
With enterprises connecting more devices and applications to their networks, IT environments are more complex than ever. The more devices you have, the more you’ll need a strong IT maintenance management strategy. Otherwise, you’ll overspend on unused devices and maintenance contracts. You’ll also put yourself at risk if an asset goes down and takes your related applications down with it.
Here are three ways to optimize your data center’s IT, so you can cut costs and reduce your risks:
1. Consolidate technologies
Consolidating your technology can give you big cost savings and efficiency improvements. For example, migrating old technology to a high-density environment can improve your cabinet space utilization by 130 percent.
If you move new technology to a high-density environment, you could fill an entire cabinet, gaining 700 percent greater cabinet space utilization and four times the processing performance. If you move to faster servers, you can save even more.
Upgrading your technology and moving to a high-density data center can bring you even greater value. For example, you can reduce your cabinets from 66 to nine, achieving capital savings of $570,000 with a cabinet consolidation of eight to one.
2. Optimize and standardize your compute environment to reduce its total cost of ownership (TCO)
Standardizing and optimizing your compute environment can reduce your hardware, software, maintenance and power/cooling TCO by up to 60%. You can save millions by conducting a comprehensive analysis of your systems and developing a strategy for workload placement.
One way to slash your software costs is by migrating any operating systems, apps and databases that are licensed on a per-CPU or per-core basis to systems with fewer, but faster, processor cores.
3. Standardize your IT to an as-a-service model
Many organizations now take a hybrid approach to internal vs. external IT services. They run mission-critical applications on-premise and their more flexible applications in the cloud.
When migrating to a hybrid model, you’ll need a partner who has experience transforming IT organizations into business enablers. They can help you navigate the complexities and interdependencies of a service-based IT model.
In some cases, using an alternate data center can help you accelerate this transformation. This gives you a clean slate to start with – free from any of your current infrastructures and processes that are hard to work around.
Pre-packaging and standardizing your IT services gives you several advantages:
Faster time to market. Rather than creating apps from scratch, you’ll already have much of your preliminary service design and integrations done. Standardization removes steps from your design process and helps you deliver apps faster.
Lower costs. Just as the assembly line allowed the auto industry to increase its margins, reusable and interchangeable parts give IT organizations economies of scale. In other words, you only pay to develop each package once.
Smoother path to automation. The old computing saying “garbage in, garbage out” doesn’t only apply to data—it also applies to processes. No amount of automation or orchestration can make up for a fundamentally poor design.
Easier management. Managing disparate systems and products is a huge drain on your time, resources and personnel.
Reduced complexity. It’s beneficial to choose strategic vendor partners and standardize with them. By standardizing servers, network, storage, security and other parts of your infrastructure, your IT organization can reduce its complexities and costs.
The scary truth about your data center
Do you know how much of your equipment may be draining your resources? Why maintain equipment that you’re not using? By consolidating technologies, standardizing and optimizing your compute environment, and moving to a standardized IT as-a-service model, you can eliminate waste in your data center. And, by taking these steps, you will also gain a deeper understanding of how your equipment ties back to the business. With freed resources and improved knowledge, your IT organization will be ready to streamline your data center and drive business transformation.