H-Connect Joomla Hosting Templateo

What if you, as a data center manager, suddenly discovered you had double the data center capacity you thought you had? 

Grossly underestimating capacity is actually much more common than you might think. One client recently showed me their data center, which they claimed was “completely full,” but they were using only half of the 2.5MW circuit they were paying for each month. Their question was why.

They explained their planning assumptions, which were very rational: 10kW racks, spec’d never to exceed 8kW. They had bought the latest in big-name blade servers, which, according to the manufacturer’s web site, were rated for a maximum power consumption of 7200W, even though each one took up just one-third of the rack. So while the rack was only one-third full, based on expected power consumption, it appeared to be completely full.

Enter a reality check. We took a duplicate server and measured actual power consumption, from boot-up to shut down, from zero-load to full-load, to get a complete picture of the server’s power consumption under real-world conditions. The most power it ever used was 3500W, and that was under full load with all memory used.

We then put in their actual use curve and found that they hit the 3500W scenario less than 1% of the time. Under “normal” use (about 99% of the time), the blade server never consumed more than 2500W.

The picture becomes much clearer now. The company could easily install two of these blade servers per rack and never hit their 8kW max. And, if they needed additional capacity, they could install three blade servers per rack and power cap them so that even under their 1% scenario—which was now two-thirds less likely to take place given the amount of capacity—they wouldn’t blow a breaker.

We then modeled the entire data center for them. They could easily double the number of blade servers installed, which would mean not only having twice the compute power, but also coming much nearer to filling the 2.5MW circuit they were paying for. And if needed they could triple their capacity and protect themselves by using the power capping capability built into their new servers.

So why is this type of over provisioning so common? The problem starts in the design phase, when typical methods for estimating needed resources virtually guarantee “hidden” capacity. Data center designers and IT managers use a reasonable process to estimate average load and peak load for an application and then add room for growth. The designers use the only sources of information on server load they have: the information the manufacturers provide. From testing hundreds of servers, we’ve found that the reported power consumption is almost always overstated (although a few types of servers actually understate power consumption, which is even more problematic).

While it’s easy in hindsight to see that over provisioning is a problem, it’s harder to justify changing this process in the absence of real information. If you don’t have information about your servers’ power usage characteristics, you can’t risk tripping the breakers. So, what’s a better approach?

Ironically, the information is all there: it just has to be gathered and analyzed.

To optimize your data center capacity, you’ll need to answer certain questions. The answers to those questions will help you use resources more efficiently to avoid systemic overcapacity.

There are three steps to optimizing your data center:

• Measure to find out how much capacity you’re really using
• Analyze to figure out possible ways forward
• Optimize to make the most of your data center

What Data Do You Need to Optimize Your Data Center?

In order to determine how much power you really need, you’ll need answers to these questions:

• How much power is your data center really using? The more granular the information about power usage, the better. How much power is being used on each rack and by each server? If you can then correlate that with your utilization on each server, you can start to understand where (and how much) power is being wasted and what you can do to change it.
• How efficient are the servers? Are the most important applications running on newer servers or on older servers that use more power and are less efficient? You might be surprised how quickly a new server can pay for itself. With Moore’s Law showing no signs of slowing down and power costs on the rise, it may make sense to upgrade more often than you think.
• How much extra load can the data center handle? Determining the true power capacity of servers as I outlined earlier gives you the real story.
• What is the total compute capacity and how is that capacity being used? More importantly, how close are you to running out of capacity? If you run out, latency increases and customers get the dreaded “server unavailable” message. You don’t want to be so efficient that you don’t have enough capacity to meet customer needs.

Data Center Analytics

The above is just one small example of how data center analytics enables you to move from guessing to knowing what you need and how to optimize it. And if you can save money in one area, you can pour it back into your company in another more strategic area that can improve the bottom line.

Data center analytics takes a holistic view of your data center. To understand how efficient your data center truly is, you need to understand not just your facilities efficiency (your PUE), but also to look at each of the following key elements and their relative capacity and efficiency: cooling, power, CPU, memory, network, storage, and space.

Whether your Web-based business is measured by your revenue per user or per transaction, knowing your detailed costs is critical to the success of your company. But do you truly know your cost per user or your cost per transaction?

The reason? The “money pit” of power costs.

Most people, if they know their total data center costs, treat power as a fixed cost. It’s pretty easy to set your budget for power. You generally know what you are going to spend each month: the bill comes in; you pay it. Over time as your user base grows and you add more equipment to your data center, your power bill goes up. But how do you assign power cost for each user, or each transaction, or, for that matter, each click? For each application or service, do you truly know how profitable (or unprofitable!) it is if you don’t know the cost of power required to deliver it?

The problem is that your servers are built to be “always on”—they are never turned off. But what’s wrong with that, you say? After all, if you are running a news service and a plane lands in the Hudson you must be able to meet the unplanned spike in demand. Whatever your business, you can’t plan for the unknown. Therefore you leave your servers on to handle the unexpected. Because of the way servers are built, however, their power consumption doesn’t change that much with application load. While your application load may vary from 5% to 95%, the power supplies in your server maintain a relatively flat rate of consumption, varying between about 70% and 90% of their maximum load. Whether it’s the first day of the quarter or the last, or whether it’s mid-day or midnight, you continue to pay your utility to keep every server running.

But what’s the cost of supporting the night owl who is paying her bills online at 2AM? How much does it cost for a 5am ATM transaction? If you could accurately measure the cost of these individual transactions, they would be much higher than transactions taking place during the peak usage times for paying bills or visiting ATMs, because you are paying for your full capacity of servers whether there are 2 million users or 2000. Power is a fixed cost, after all. The key question is: does it have to be?

What if your actual power use went up and down along with the number of users? If your power draw could vary between 5% and 95%, just as your application load does, you could assign power costs accurately to each transaction (or click) and know your true cost per user or transaction.

Power Assure helps you determine your true costs. By “virtualizing” power, we in effect create a flexible pool of servers that turn on and off automatically along with your application load. Because power now increases or decreases with customer demand, it is relatively straightforward to assign the amount of power each user or each transaction is consuming. This gives you a much clearer picture of the true cost of an application, and therefore, its profitability.

At any time of the day, Power Assure specifies the exact number of servers you need, plus a desired safety buffer to be able to handle quick unexpected spikes. And whether demand hits the roof or drops to nothing, your power costs will go right along for the ride.

Following up on the topic of Power Virtualization, “virtualizing” power has multiple benefits, especially if you are running more than one data center.

By spreading the load among various sites, power virtualization allows you the flexibility to move computing on the fly. In response to an outage,

However, there are many advantages of being able to move compute resources quickly to wherever needed. Power Assure allows you to make use of variables other than service availability upon which to base the decision to move processing to a different location.

Take price, for example. The amount you pay for power varies by time of day, by season, and by location. If you are running a data center in midtown Manhattan during the peak demand hours of 10AM-2PM, the price you pay could be many times what you would pay in North Dakota at the same time. By turning on price as a variable in load-shifting, Power Assure can instantly move your primary data processing from Manhattan to North Dakota (or wherever you want it to go) while prices are high, allowing you to save, literally, millions in annual power costs. The servers in Manhattan would remain in the pool waiting to be turned on if needed to respond to an outage at another center, or a spike in demand that exceeds the combined capacity of your other centers. And when the day is done the Manhattan servers are turned back on and can then take on capacity from another data center where costs may be higher later in the day. And if latency is a concern, you can set limits, e.g. no shifting during trading hours.

Yet another variable is Demand Response (DR). Many companies are signing up with providers like Enernoc, Comverge or CPowered in order to take advantage of the millions of dollars utilities are willing to pay if your company can reduce power at a specific site very quickly. The way it works is quite simple. The utility sends a message asking for power reduction; if the site can support it, Power Assure automatically shifts the load to other sites, temporarily reducing power consumption at the DR request site. Once the DR period expires, Power Assure can transfer the load back. Without any sacrifice in service quality, you sit back, let the software do its work and wait for your rebate check.

“Virtualization” is the hot topic in data centers today. You can now virtualize your servers, allowing your CPU and memory resources to be treated as a single pool. CPU virtualization has allowed companies to run multiple applications on a single server and thereby more efficiently utilize data center server capacity. In a similar way,

But there is yet a third virtualization opportunity that companies are just beginning to exploit: Power Consumption

Currently, power consumption at any data center is essentially fixed. However, imagine a world where all power consumption at all of your data centers represents one large pool of power consumers. What if you could increase or decrease the size of this pool to match customer demand? Or, if you have multiple data centers, what if you could move your power consumption around from one data center to another as demand increases, or as power prices change by time of day? As a result, power consumption essentially becomes virtualized, since power consumption at any particular data center is no longer fixed.

Impossible? No. In fact the technology to make this happen is available today, and deploying it can cut the cost of power in your data centers in half. So why does the “virtualization” of power consumption save so much?

Most data centers are built for peak load. And because of this, all of the gear—the servers, the switches, the routers—are left on all the time. This “always on” model assures that you will always meet any spikes in customer demand. But the problem is that it costs a fortune. Power consumes 25% - 30% of the average IT budget. U.S. companies spend more than $10B to power their data centers each year. With this “always on” model, power has essentially become a fixed cost.

The problem comes from the way servers are designed. Even if your CPU utilization varies from 5% to 95%, the amount of power the server draws only varies between 70% and 95%. So while your actual user demand may vary dramatically over the course of a day or a week, your power use actually changes very little. In fact, it really goes only one way: up! The more capacity you add, the higher your power bill.

Power Assure allows you to treat your servers—whether in one data center or across many—as a single pool. We calculate the exact number of servers required at any time to meet your actual user demand and keep these machines running at full capacity. We shut down (or put to sleep, whichever you prefer) the rest, and put them into a pool of “available capacity”. As demand increases, we turn on however many servers are needed to meet the spike and support the customer demand. All of this is done dynamically, in real time.

The end result is that power consumption becomes virtualized. You only pay for what you need to meet customer demand at any point in time. Your servers are treated as a single pool—or as multiple pools if you are running multiple applications—and the resources are distributed according to need and location. Just as with CPU and storage virtualization, the result is increased flexibility, higher efficiency, and significantly lower costs.