Storage Wars Expand With Google Ready to Compete With Carbonite and Dropbox
by Mark Mills
February 24, 2012
One of comedian George Carlin’s great riffs was about our propensity to cherish and store “stuff.” We all like to store stuff.
So it’s a big deal when we learn that Google will soon launch a consumer-oriented Cloud storage service called Drive. Another place to store our digital stuff.
Google is highly attuned to what its customers’ want, and quick to recognize emerging opportunities. The entrance of the Big Dog of the Internet into the rapidly growing Big Data storage business is a huge deal. And competition aside, it’s a promising bellwether for the existing family of storage-only companies like Dropbox (IPO soon?) and Carbonite [NASDAQ:CARB].
The underlying behavioral macro here goes beyond technology. Permit a brief digression into relevant consumer behavior in the self-storage industry where billions are already being spent to fill warehouses with physical stuff, and what it says about the billions that will be spent to fill data warehouses chock-a-block full of gigabytes of cherished virtual stuff.
Over the past three decades, America’s boxes-and-sofas self-storage industry – comprised of companies like Public Storage [NASDAQ:PSG] and U-Store-It [NYSE:YSI] — has grown ten-fold to a $22 billion a year business. In fact, it’s been the fastest growing sector in commercial real estate, with five public companies and hundreds of smaller operators. Demographics and wealth drove this trend; everyone has more stuff. Well, at least they bought a lot of stuff while the economy was expanding.
Once upon a time, a century ago, people didn’t need to store much stuff away from home because no one had much of anything. Never mind off-site, most houses didn’t even have or need much in the way of closets. The huge drop in the real cost of things since 1912 led to people owning a lot of everything; there’s been a roughly ten-fold drop in the real cost of men’s suits, women’s shoes, golf clubs and even books. More stuff. Gotta put it somewhere … to paraphrase Carlin. The self-storage industry manages two billion square feet of space now; enough space to store about 30 books or photo albums for every American.
Once upon a time, an Internet ‘century’ ago – say circa 1985 – people didn’t need to store digital files. They didn’t have any. And when they started to get some, they were expensive…you know the story.
Now it’s nearly trivial to store 30 photo albums worth of digital pictures, or the video equivalent, for every American. But so far the self-storage industry for consumer data hasn’t broken the $1 billion revenue mark, globally. But with data storage growing at nearly 50 percent a year, it will only take a little over a half-dozen years for consumer virtual storage to become a bigger industry than physical storage.
In an intriguing and temporary coincidence, storing a physical book in a U-Store-It warehouse today costs almost exactly the same as storing a virtual book in the Cloud — about $0.20 per year per book. Pretty obvious which is more convenient.
The future cost of storing physical things is essentially fixed, tied to slow-moving metrics like the cost of land, steel and concrete, and labor. The future cost of storing digital things is still collapsing at Moore’s Law kind of rates – which is to say geometrically.
For the energy conscious amongst us, a related interesting factoid; storing a book virtually also costs about the same today in terms of just the electricity used as it does storing its physical counterpart in a warehouse. And for efficiency mavens, unlike the ‘real’ world, the energy used per book stored in the virtual world will keep falling. But the universe of book-like things to be stored virtually is poised to explode at a vastly higher rate.
At the dawn of the Internet – circa late 1980s – it cost $10,000 for the electronic hardware (disk drives) to store the content of a single book. Today that amount of storage hardware costs just two cents. It would be one thing if just the costs of storage were dropping. It’s quite another if the supply of stuff you want to store is also rising at a similar geometric rate.
Begin with the conventional wisdom. About 1.5 billion people globally are connected to the Internet today. The biggest self-storage data-warehouse players so far – Carbonite, and Mozy, a division of VMware [NYSE:VMW]– can count only several million customers so far. Not even a scratch on the digital surface. It’s reasonable to think that 10 percent of consumers will eventually want and need data storage (the same market share as the physical self-storage industry) – that yields 150 million more users.
Going forward? The connected universe can only grow. That’s what underlies Google’s drive to create Drive, Apple’s iCloud [NASDAQ:AAPL], and Amazon’s Web Service [NASDAQ:AMZN] to provide rentable Cloud infrastructure – a virtual U-Store-It on steroids.
By most forecasts, at least another billion people in the world will join the Internet in a few years. That’s just the beginning. It’s not just that more and more people will connect, but many will increasingly have more than one internet-connected device. Over a billion data-creating devices are sold each year now – counting desktops, laptops, tablets, smartphones.
Add to this a broader context. A recent OECD report on Machine-to-Machine Communications puts the number at 50 billion mobile wireless devices, with another 500 billion devices connecting to the Internet by 2020. Read the report. While much of this universe is corporate (factory automation, supply chain logistics), a lot of these devices are consumer-related, from eHealth to smart automotive and home features. Stuff we care about. Data we own. Stuff we just might want to store ourselves, directly, or indirectly.
Last week we saw a hint of the trend when Amazon Web Services reported a nearly 200 percent growth in objects stored year-over-year. At those growth rates virtual storage revenues blow past sofa-and-box storage in just a few years.
Consumer-oriented digital products and services have a velocity all their own. Thus the excitement over last summer’s IPO of Carbonite was more than a break from the paucity of recent IPOs. Analysts like those at Bank of America Merrill Lynch and Canaccord-Genuity covering CARB are bullish. Understandably. Of course there are other players in the consumer-facing part of this market including: ElephantDrive, Fastmule, Keepit, LiveDrive, SugarSync, Backblaze and JustCloud. (By the way, for thorough testing and ratings for Cloud back-up see CloudBackUping.) The emerging market is big enough to absorb a lot of entrants.
For Carbonite to stay a leader, the key is to have enterprise software clever enough to get the highest utilization, efficiency and speed out of the storage hardware, with an easy, consumer-friendly user-interface. They’ll have to keep innovating, of course. The battle for mindshare won’t be won just with good advertisements and celebrity endorsements where Carbonite is a clear leader, but with good user experience. The reports from the field are encouraging.
So on one side of the ecosystem we have the Carbonite-class companies bringing ever-better interface and user experience, sitting on top of the infrastructure that makes storing gigabytes not just easy, but trivial. At the other end of the ecosystem resides the tech community that is driving the core trend in storage infrastructure – driving down the costs of storage. In the digital world, costs drop because speed and efficiency get better. Cheaper, faster storage is the rocket fuel for companies at the pointy end of the stick (as my military buddies like to say). How much cheaper can storage get? A lot.
Dynamic (i.e., as opposed to read-only optical disks like DVDs) digital data are stored almost entirely in one of three types of hardware; tape drives, hard drives, and solid-state or flash drives. All three are getting cheaper and expanding in the new zetabyte era, including tape. We’re not talking about your pappy’s 8-Track tape from a ’72 Buick.
For archiving it’s hard to beat tape. The more data and the more valuable the data, the more important archiving becomes. Compared to a hard drive, tape is a tiny fraction of the cost, uses nearly no energy when idle (literally, put a spool on a shelf), and lasts decades without degradation. All great for archiving. But tape’s speed is not even close enough for the day-to-day life at the microsecond level to transport terabytes in real-Internet-time inside a data center.
Rapidly spinning hard drives have anchored both consumer computing and data centers for years now. According to Gartner, total hard drive shipments have grown from about 100 million terabytes five years ago to nearly 1,000 million terabytes forecast this year. The collapsing costs of hard-drive technology can be credited with fueling storage growth and enabling stored-and-streaming video, as well as companies like Facebook.
The next leap in storage architecture for data centers comes from the same technology that has made the MacBook Air, and the family of imitators, so popular now – flash drive, or solid-state drive. Microchip memory instead of spinning disks. Blazingly fast. Low power. Solid-state-drive-based computers turn on nearly instantly and have longer battery life from lower power drain. Put flash memory features into data centers, and Katy-bar-the-door. That’s what a new company called Violin Memory is doing.
Violin’s flash-based hardware, combined with clever software to take advantage of the blazing speech of flash in huge arrays, enables breathtaking gains in data-center storage capability. Their credible claims are for increases in speed measured in thousands of percent, and at a fraction of the floor space and power requirements. These are the kinds of capabilities that create a vortex of demand. Storage costs will radically collapse and light a fire under the capabilities of storage services across the Cloud, including and especially for consumers. One would expect an IPO in Violin’s future. Stand by.
There is yet more technology on the horizon to push the speed and cost envelope even further. Perhaps most tantalizing will be the eventual commercialization of either IBM’s new nano-tech that blends the attributes of magnetic hard drives and solid-state memory, called Racetrack memory, or HP’s radical low-energy memristor. While it will take time, likely both will emerge in due course as well.
The overall domain of all data stored for the public and private Internets is already much bigger than the nascent consumer-facing business. The 15 petabytes of new data created daily (it was only a 1995 that the entire Internet barely approach one petabyte a year) makes for unprecedented storage challenges within the walls of data centers, from access speed, to reliability, to sheer power and cooling needs. No surprise last year HP [NYSE:HPQ] and Dell [NASDAQ:DELL] sparred to buy 3PAR and its storage optimization for data centers, with HP paying $2.4 billion for the $100 million business. Those kinds of acquisition multiples reflect a firestorm of activity across the entire storage ecosystem.
This huge technological infrastructure is why low-cost consumer-oriented storage has been enabled … we’re surfing behind a massive bow-wake of storage technology progress and infrastructure. New businesses and techniques are being developed to deal with scales of data storage unprecedented in human history.
One of the most intriguing challenges at the data-center and enterprise level is just keeping track of the overwhelming volumes of files. There are a number of new players emerging to address this, including FileTrek, a particularly compelling start-up coming out of stealth mode only recently. (And they’re based in Ottawa, Canada, my hometown.)
Also watch for emergent plays in analytics to extract intelligence and trends from Big Data – that’s what Microsoft appears focused on, and IBM was chasing in building their Watson non-linear computer (see my earlier column on Watson). Before long, we’ll all be using Watson-like capabilities to mine our personal data. Maybe a feature Carbonite or its competitors will add in the future?
Storing digital stuff in the Cloud is set to be one of the biggest and fastest growing infrastructure and investment trends of the decade. Expect the battle ground for mindshare and market share to be intense, both on the consumer service front lines, and back in the trenches with enabling technologies – the ‘ammunition’.
And to quote the inimitable late George Carlin, paraphrasing the 4th wisdom in a list of 101 greatest Carlin aphorisms:
“A [data center] is just a place to keep your stuff while you go out and get more stuff.”