Photo: Peter McCollough / Wired
In 1956, the first modern container ship traveled from Newark to Houston. It seems like such a simple idea now: a boat carrying not random little crates, but massive containers that could easily be moved onto trains and trucks and straight into factories. The arrangement took years to construct, perfect and implement. But once it was put into use, the global shipping industry would never be the same.
Containers revolutionized the shipping industry because they made it easier to transport massive quantities of inexpensive goods from port to port. Suddenly, small crates – and the shipmen who earned a living loading them on and off boats – became irrelevant. Goods could stay in their containers for the duration of their journey from one factory to another, saving significant time, money and effort.
Today, an analogous transformation is taking place: This time, we’re transporting digital, not physical goods. Rather than physical factories, we’re seeing the rise of digital ones.
Or, as they’re more commonly known: data centers.
In this analogy, those small crates and hardworking shipmen are the former powerhouse data-center server makers Dell, HP and IBM. Whereas businesses once purchased servers from these big three to store and transport their digital goods, they’re now choosing to make their own servers and data centers.
Why? Because there’s greater potential for efficiency. The rise of the digital “data factory” will be similar to the rise of the physical factory in how it affects different businesses. Namely, big businesses will build their own. And those little “crates”? They’ll become a thing of the past.
We’re already seeing this revolution in motion. The heavy hitters in tech – think Facebook, Google and Amazon – can’t survive on small-scale servers. So they’ve gone on to build out their own “data factories” that are faster, more scalable and more flexible. These self-made data centers can process a lot more data than the standard packaged ones with which server vendors have long dominated the market.
Further evidence of this can be seen from some interesting data Intel CIO Diane Bryant shared with Wired just a few months ago. She said that in 2008, 75 percent of Intel server chip revenue came from three server makers: IBM, Dell and HP. In 2012, it’s eight companies, not three, which account for 75 percent of revenue.
Four years ago, most companies were buying servers from the big three – but the new statistics from Bryant indicate there’s now been a shift in the market. HP denied the validity of these numbers, stating that HP, Dell and IBM combined still held 73.9 percent of market share. But if you look at the how the data center has actually evolved, Bryant’s data confirms the server industry is being disrupted.
And three key trends have evolved as a result of this disruption:
Individual companies have grown as big as the big server vendors. Today, there’s so much data in existence that a single company can command a server volume close to the size of a single vendor’s entire business. If Dell has a volume of 2 million servers and your company requires 1 million, you suddenly comprise half the total volume of one vendor’s business – and that’s when it stops making sense to do business with the middleman. Just ask Google, which sits pretty at number five on Intel’s list of top eight server makers. The search giant doesn’t even sell servers; it simply needs so many of them to run its business that it makes more sense to build the servers and data centers themselves.
Companies making parts are consolidating and standardizing. Another trend we’ve seen is fewer companies in the business of making server components. We’re left with just three major fabricators in Intel, Samsung and TSMC; and just two major companies that make hard drives (Seagate and Western Digital). This unprecedented consolidation makes it significantly easier for companies that want to build their own data centers, because it means more standardization of parts. Companies can pick and choose the parts they want to use from all three fabricators.
Components are becoming commoditized. Once, the servers themselves were the commodity. But now, we’re seeing the commoditization of server parts. The purchase of these server components – boards, network interface controllers and chips – signifies a shift in the market. It indicates that companies are making their own servers and data centers, rather than buying hardware from someone else and renting space in someone else’s computing facility. And perhaps more significantly, Intel has taken note of this trend and has started offering warranties on individual parts.
What we’re seeing here in the world of servers and data centers is exactly what we saw when factories and containers revolutionized the shipping industry over 50 years ago.
Shipping containers changed the way we store and transport physical goods. Modern data centers change the way we store and transport digital goods. Computers and servers were the focus of our industry before the data center, just as small crates were the focus of the shipping industry before the container. But now data centers are being built and customized … just like the containers and factories before them.
What does all this mean? It means big businesses will build and manage their own data centers. Small businesses will want to rent space in the bigger data centers.
It’s a new sort of “cloud computing”; a holistic concept that will maximize efficiency and take our economy to the next level. But the bottom line is that this business will no longer be about who’s building the box.






![Creative: Muhammad Suhaib
After two years, Millat Facebook is on the verge of collapse. By Rana Tanveer
Anti-western anger may bring out street power, but, at least in Pakistan, does not appear to be commercially viable. MillatFacebook, meant to be a religio-nationalistic alternative to the global social network Facebook, is in dire financial straits and is seeking donations to pay its server hosting bills.
In an e-mail sent out to its users on December 10, MillatFacebook appealed for donations to pay its servers, which it says are based in the United States. Yet the e-mail was riddled with factual errors and highly suspect claims, casting doubts on the integrity and intellectual capacity of its owners.
The website, for instance, insists on calling itself “the only Facebook rival”, ignoring the existence of a wide range of social networks such as Google Plus, LinkedIn, and others. The e-mail starts off with a person introducing themselves only as “S.O.A.” and then launching into a poorly worded e-mail appealing for funds. In the e-mail, the company claims to be spending thousands of dollars on its web hosting, and then says that it is looking to raise $581 from its users in order to defray those costs.
“As you can imagine, cost of running a social network is never easy and very costly. Current month we are facing a shortfall of 581USD (sic) in server payments. (We are paying 1000s of USD in server payments ourselves and now need your helping hand to meet this shortfall). We are asking for a contribution from you that will help us towards our goal of keeping this Peaceful Social Network run forever and defeat blasphemer facebook. inshaALLAH, Aameen. Any amount that you can contribute would be greatly appreciated,” said the e-mail.
The claim that the website requires thousands of dollars in web-hosting costs seems dubious. According to websitetrafficspy.com, a leading source of internet traffic information, millatfacebook.com attracts an average of two users a day. With traffic that low, it seems unlikely that MillatFacebook would need to buy expensive web hosting facilities.
The website is asking its donors to give money through credit card payments on its homepage, but Chief Operating Officer Umar Zaheer Meer admitted to The Express Tribune that the fundraising drive was not going well. He claimed that donors were reluctant to give money via credit cards owing to “cyber laws” which he did not specify.
According to Meer, MillatFacebook has over 476,000 members, though he did not specify how many were active. When it was launched, MillatFacebook claimed that it would not have any paid advertising, though it now has ads and still seems to be struggling to pay its bills. By contrast, the Silicon Valley-based Facebook has over one billion active users as of October 2012 and had annual revenues exceeding $3.7 billion in 2011. According to a recent report, more than eight million Pakistanis use Facebook.
The website was launched in May 25, 2011, six days after the Lahore High Court issued an order banning Facebook, after one of the website’s users created a competition called “Everyone Drawn Muhammad [PBUH] Day”. Muhammad Azhar Siddique, the man who filed the petition that resulted in that temporary ban, convinced Omer Zaheer Meer, now CEO of MillatFacebook, to start the website. The hope was that the ban on Facebook would be permanent, forcing many Pakistanis to look for an alternative.
The website was launched at a ceremony at the Lahore High Court Bar Association and was billed as the alternative to Facebook. Yet, unluckily for the founders, access to Facebook was restored on May 30, 2010, snuffing out any chance the website may have had to catch on amongst Pakistani internet users.
The Meers appear to despise Facebook, yet have designed MillatFacebook to look almost exactly like their supposed rival. Users of both websites, however, say that MillatFacebook is a far less functional website, with one user going so far as to say that “the quality of user experience is so abysmal that it does not merit the humble title, ‘Facebook clone’.”
MillatFacebook is not even the first social network targeting Muslims. Naseeb.com, also based in Lahore, was launched in 2003 and precedes even Facebook.
Published in The Express Tribune, January 2nd, 2013.](http://25.media.tumblr.com/260a2dcc7831c2108ad56943128f65d8/tumblr_mg22f3FICF1qeoyxwo1_500.jpg)





Getty
Getty

