The simple silicon on which so many data centres are based has spent decades as the information workhorse, but it’s starting to show its age.
There is little chance that it will be able to keep up with the imminent flood of data driven by the 26 billion connected devices that Gartner predicts will be in use by 2020.
Add that to some 7.3 billion smartphones, tablets and PCs that will also be active by that date, and you can begin to imagine the volume of data that we’ll be dealing with.
Most estimates place the total amount of data that will be generated in 2020 at over ten times the amount produced in 2013: a jump from about 3.5 zettabytes (that’s 35 with 20 zeros following) to 44 zettabytes of data annually.
>See also: CIO as Frankenstein: Creating a monster infrastructure with web-scale IT
Breaking these numbers down, 44 zettabytes would roughly equate to one million photographs or 1500 HD movies for every single person on the planet.
At the moment, the approaching gap between our data output and our storage capability is not a problem that can easily be resolved.
Building factory capacity that is capable of matching the scale of such spectacular demand would take hundreds of billions in investment.
Even if these stratospheric costs could be met, the fact of the matter is that it’s also far harder to manufacture zettabytes of capacity than it is to generate zettabytes of data.
Ultimately, it’s becoming increasingly difficult on a molecular level to squeeze growing volumes of information onto the same amount of space.
Since the first hard drives in the 1970s, the storage industry has roughly conformed to Moore’s Law – every two years hard drives have approximately doubled in capacity and halved in price, without getting much larger in physical size.
A 1979 device for example had 5MB of storage, cost a few months’ wages and could save about two seconds of low resolution video shot on one of today’s smartphones, or two photos.
A 5TB hard drive today costs less that £200, can store 2 million smartphone photos, 2.5 million songs and about 1,000 movies. Although it’s not any bigger than the 1979 device, in capacity it’s actually 1,000,000 times larger.
If similar advances had been made in engine fuel economy, we’d now be able to drive to Jupiter and back on a single tack of petrol.
While a testament to human ingenuity, writing data in increasingly dense layers on silicon sheets without adding additional weight and bulk to the device is no longer the solution to our device dependence.
Fortunately, a demand that is not being met by production is certain to spark the invention of new technologies and we are certainly starting to see impressive innovation in the field.
From the breakthroughs in the efficient production of RRAM, to work on HAMR technology and the mind-boggling complexities of DNA storage, there is much work being done to extend the life of silicon as much as possible, and to imagine entirely new methods of digital storage.
With these innovations and those sure to come, it may be possible that the world will be able to close the data capacity gap entirely by 2025.
In the meantime, businesses and CIOs should be factoring storage into both long and short-term plans as these data guardians will soon be facing a Catch-22 situation.
Data is quickly becoming as valuable as oil but in order to amass it and keep it safe there must be ready capacity.
In order to drive the most value from stored data, it must also be saved in such a way as to be quick to access, efficient to manage, and low-cost to maintain – therein lies the rub.
Data centres today are not equipped to be able to handle the anticipated influx, nor geared towards feeding it smoothly across to the analytics platforms where it can prove its worth.
They tend to be built around legacy components both in terms of hardware and software stack components.
The result of this unwieldy, legacy approach is inefficiency. For example, in a single system there can be several components caching the same data, each working independently and out of sync, with very poor results.
In order for a company to get to a better cost and efficiency model, so as to better match future requirements, another solution must be put in place.
One potential solution is a tiered model of storage. SSHD (solid state hybrid drive) hybrid technology has been used successfully in laptops and desktop computers for years but it is only now that experts are beginning to consider it for enterprise-scale data centres.
This model uses a more efficient capacity-tier based on pure object storage at the drive level, and above this a combination of high performance HDD (hard disk drives), SSHD (solid state hybrid) and SSD (solid state drives).
>See also: How to avoid becoming a big data liability
The combination of storage drives allows the most critical data to sit on the more expensive SSDs or hybrids, where it is quick and easy to access.
The less important bulk data sits on lower-cost HDDs where it is still available and secure, but slower to process.
Today we are only in the very early stages of expanding this new data storage method to business, however it’s a smart response to big data requirements and one that I expect will be widely implemented over the next two years.
Good business the world over is reliant on maximum productivity, which requires speed, efficiency and process management.
Taking time now to make smart storage plans for the future could be the difference between the market leaders of today and tomorrow.
Sourced from Mark Whitby, Seagate