SPONSORED: Triple-Tiered Storage Makes Sense
SPONSORED: Triple-Tiered Storage Makes SenseEffective, cost-efficient storage solutions demand a multi-tiered approach, with different types of drives and controllers for different application needs.
Nobody drives a racecar to haul hay, and no one drives a tractor on the racing circuit. Each vehicle has its own specialization, application, and price band. You pick the right car for a given task. Data storage adheres to the same principle. We pick different types of storage for different needs.
Enterprises must process a fantastic amount of information, and the need to store much of it is only increasing as customers want limitless access to historical content and companies demand ever more bits for mining and reselling. The amount of data needing to be stored continues to rise exponentially. Science magazine recently published a report (Science, 1 April 2011, pp. 60-65) by Hilbert and Lopez claiming that the global capacity for digital data storage in 2007 was 276 exabytes. (An exabyte is one billion gigabytes.) Less than four years later, a Cisco white paper titled “Cisco Visual Networking Index: Global Mobile Data Traffic Forecast Update, 2010-2015” illustrates how mobile traffic alone will soon account for several exabytes of data per month. By 2015, 66% of this data will be video, and, if today’s usage models are any indication, much of that video will be stored—multiple times.

If Cisco’s projected growth rate holds, just the mobile traffic generated each month in 2022 will exceed the world’s total digital storage capacity of 2007. Certainly, some of this will be dancing baby videos, but there is also the opposite end of the spectrum: banking data, encrypted health records, point-of-sale databases, and the like. If it takes 15 seconds to bring up a dancing baby video, few would mind. But 15 seconds to respond to a client’s file request is an eon, and long-term it’s a business-killer. Some data simply must be accessible from storage on a practically instantaneous basis, and other data can sit in slower, less expensive repositories. You can think of this in terms of temperature, with the fastest storage holding the most critical data being “hot” and the storage containing the slowest, least important data being “cold.”
“If you’re an airline and somebody wants to book a flight, being able to get those schedules out and move those transactions quickly is going to make those customers happy, and it’s going to increase the likelihood of getting revenue,” says Seagate Senior Product Marketing Manager Teresa Worth. “Once a flight is booked and the airline has payment, then that itinerary data might move into or a warm data classification. The customer still wants quick access to it if they want to change the flight, check seats, or whatever, but it’s not ultra-hot anymore because the airline has already closed the sale. After the flight happens, then that data can go into cold, or nearline, storage. It needs to stay accessible because maybe the user wants to match up mileage points at the end of the year and make sure they got credit, but a little wait isn’t going to be a problem.”
As with cooking, temperature is everything. Perhaps also like cooking, the hot materials get most of the attention, but it’s the cold stuff, the items sitting in the refrigerator and freezer, that consume the most volume. Hot storage tends to be on low-capacity drives; cold storage drives can generally hold several times more data. This dynamic sets up certain trade-offs that must be considered when designing tiered storage. The strategies with which a company segregates its storage into hierarchical tiers can have a significant impact on the business’s bottom line and, in some cases, even its life or death.