Posted by: Greg Ness | April 1, 2011

Welcome to the Era of Mass Computing

There is a vast new technology infrastructure now building itself out with significant implications for Archimedius readers.  As you might recall from previous blogs, in 1999 -in the midst of the “dotcom boom- there were fewer than 100 million IP addresses.  We are in the midst of a new computing revolution involving the mass adoption of computing technology and the dotcom adoption pales in comparison.  Hundreds of millions of IP addresses are now being added each and every year. Archimedius readers who understand the depth and breadth of the implications of this accelerated adoption of computing technology are more likely to benefit from yet another technology disruption. 

The investing advantage often goes who know where to look: technology-disruption today is very likely to happen in areas where mass adoption creates new demands; areas like specialty silicon (endpoint processors and touchscreens), infrastructure 2.0 networks, virtualization and cloud (IT automation and optimization) and wholesale data centers (the mass computing era wheelhouses) are all worth reviewing.

My apologies for quoting the Intel CEO yet again, but I think many of us are so numbstruck by the pace of innovation that the sheer magnitude of what Paul Otellini has said (in Investing Daily) has yet to sink in:

In 2010, total traffic crossing the internet was roughly 245 exabytes. This is greater than all the previous years combined. Over the next five years a billion more people will join the global online community, with 15 billion new connected devices including PCs, smartphones, tablets, embedded devices and smart TVs.

We estimate this will increase the data footprint across the Internet to over 1,000 exabytes, more people, more devices, more usages. This dynamic will require high-performing servers from Intel for years to come.

Perhaps the dotcom boom has taken the bloom off the disruption rose and we’re so afraid to acknowledge the sheer implications of the mass adoption of computing that we’re distracting ourselves with one of the several mere undertones, like cloud computing.  Yes, cloud is an operating model subset of a greater mass adoption of computing This mass adoption is a far greater story than cloud.

Mass computing may turn the tables on status quos and forever change the economics and importance of the IT industry, in the same way as the automobile and the train and the airplane drove the petroleum and transportation and shipping industries to new heights and redistributed wealth on massive scales.  There is an interesting array of new benefactors, depending on their ability to execute on their vision; their futures may well depend on the level and pace of new demands on devices, networks, servers as well as the traditional data center itself.

Specialty silicon and sensors (for these exploding populations of endpoints) would be a natural consideration, including companies like Cypress Semiconductor (CY) and Atmel (ATML).  When you take a look at the group look at who they are supplying; they win or lose based on the success of the devices they’re shipping in. 

Companies which help to automate network functions (the infrastructure 2.0 gang  of Cisco (CSCO), F5 Networks (FFIV) and Infoblox now also including Juniper (JNPR), Riverbed (RVBD), HP (HPQ) and Arista Networks,  etc. because of recent announcements regarding network automation and management capabilities) and manage ever-increasing network-attached appliances also offer interesting upsides if they’re able to help enterprise IT innovate out of the rising manual labor demands on increasingly taxed networks.

Think about the economics of delivering applications from servers instead of hard drives to ever-increasing populations of users.  Processing demands shift from motherboards to ever more powerful servers; this is the opposite of the previous trend to ever larger software apps on ever larger hard drives that fueled the PC boom and sent valuations of companies like Microsoft and Intel into the stratosphere.

The virtualization players, including VMware (VMW), Citrix (CTXS) and Microsoft (MSFT) are interesting in that they can increase the operating efficiency of IT departments, at the time that they are facing new demands and higher levels of financial scrutiny.  More racks and stacks equal higher costs which translates into more scrutiny.  They are racing to deliver ever increasing management, security and flexibility capabilities as they race to convert legacy environments to ever more dense and powerful blade servers housing hundreds of virtual ones.

Cloud computing is so amorphous that I’ll merely “tip my hat” to the application, infrastructure and platform as a service plays and let the high profile scrum continue.  I think it’s a strategic subset of this greater move to mass adoption of computing technology.  We’re witnessing a whole new renaissance in the service provider industry as a result of virtualization and cloud, as well as the rise of the big “O”: opex, or operating expense as a strategic consideration. 

You can also read more about networking, virtualization and cloud here and at Archimedius.

This is why I’m also very interested in the emerging wholesale data center space; so interested that I joined a player in the space (more later- remember Victor Kiam?).  A variety of analysts have documented the tired state of the enterprise data center, with most data centers today being more than seven years old.  Yes, many were built when the hard drive was the primary mechanism for delivering endpoint software, or hundreds of millions of IP addresses ago, before the rise of social media and the mass adoption of online gaming, not to mention the rise of cloud and virtualization.  These data centers are very inefficient for the most part and yet represent a substantial amount of domestic power consumption in modern economies.

You can read more about the trends driving more efficient data centers from a recent Pike Research announcement.

Players like Digital Realty Trust (DLR), CoreSite (COR) and Dupont Fabros (DFT) and (my employer), privately-held Vantage Data Centers, are well positioned to lease wholesale data center space to enterprises who want the operating benefits of sizable capital investments in design and location/amenities without the upfront capital costs.  As some enterprises go it alone and build data centers as one offs, the wholesale players can introduce unique arrays of advantage to their competitors, often with less time and risk. 

As IT becomes ever more strategic, operating advantages can increasingly make the difference between success and failure.  This will shift advantage from real estate-centric wholesale players to the engineering-centric.

These unprecedented demands on servers, for example, are likely to drive even more demands for vertical scalability, efficiency and flexibility, which would mean substantial reductions in IT operating expenses.  Vertical scalability allows enterprises to turn up the dial on server power and cooling without tearing down walls or having to build/buy a new building.  Vertical scalability is a powerful new data center capability, and has thusly become a catch phrase (like cloud) for the ability to temporarily add more square feet or power- that is, until someone else needs it.  True vertical scalability isn’t dependent upon additional space or the availability of other tenant’s power. 

Efficiency translates into more of the available power being used for computing infrastructure versus being lost in the building.  The more efficient a data center the more of the consumed power that actually reaches the IT infrastructure.  Highly efficient buildings can thusly reduce IT operating costs by millions of dollars per year by minimizing power and cooling costs.

Flexibility is the ability of a provider to address unique enterprise needs, whether its room temperature, scale needed, redundancy or any of a host of specific requirements.  Thus far the industry has not been flexible enough for most enterprises, hence the interesting upside to the category if true flexibility is embraced.  Again, this is why the industry is likely to shift from being driven by real estate to engineering.

There are no doubt a multitude of other categories that stand to benefit from the shift to mass computing, including data security, IT management and orchestration, and new kinds of service providers, from Apple (AAPL) and Amazon (AMZN) to social media and online gaming players.  This is my scan of the horizon based on my industry perspective and innumerable coffees and lunches and tennis matches with Silicon Valley colleagues.


Leave a Reply

Please log in using one of these methods to post your comment: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s


%d bloggers like this: