Technology, journalism, social media and social responsibility
Intel’s announcement that it intends to purchase McAfee, Inc. for a record-breaking $7.68 billion came as a surprise to many industry watchers. Indeed, there are about as many theories about why the hookup took place as there are viruses on the Internet.
One suggestion is that Intel needs a better hook for its problematic portable chip business, which is losing ground to competitors who can produce a lower power device that is far more attractive to the market. But the addition of a security layer, no matter how good, isn’t really a differentiation that most people would expect to see at the chip level; typically, security software resides as a stack on top of the chip’s firmware, where it is easier to update and modify. And it doesn’t solve the power problem, which is being addressed by Intel in other ways.
Another theory that has been suggested by a Wall Street analyst says that Intel is trying to become more like IBM, branching into the service business. But this theory leans heavily on the proposition that Moore’s Law is in fact coming to its final physical limits – something which Intel has refused to acknowledge directly, although increasingly, Intel executives have been suggesting that Moore’s Law is an economic model as much as a physical model, thereby opening the possibility there are other ways to achieve the desired cost savings in addition to reducing the linewidth of semiconductor transistors.
This is good stuff, so let’s hold on to it for a second while we consider yet another factor, a challenge that is facing both Intel and all of its competitors: the aging American telecommunications infrastructure.
The US telecommunications grid is woefully inadequate to support the amount of digital traffic that is being pushed through its pipes today, let alone the amount that will be required five or 10 years from now as mobile devices become orders of magnitude more powerful than today’s desktops.
The idea of completely replacing the infrastructure is untenable, especially given our nation’s present economic condition. Nobody wants to do it, and even hookups such as that between Google and Verizon don’t really address the issue head-on.
One way companies have been trying to overcome this problem is to push the data out so that it is always closer to where it is needed, thereby reducing the strain on the system. So for example rather than having huge centralized data centers, we could have many smaller distributed data centers in strategic locations around the globe, each carrying a slightly different data package depending on the requirements of the region. This is the basic premise of cloud computing.
This concept can be replicated on a smaller scale by local ISPs or corporate data centers, which can store information that is most likely to be required by local customers. WiFi and WiMax services are another incarnation of this idea, delivering “last mile” and smaller broadband connectivity. Rather than providing phone lines for every user in a location, the company installs a WiFi or WiMax system, often with localized data access. Hospitals are becoming huge users of this capability as it saves hundreds of thousands of dollars that might otherwise be spend pulling cable all over their buildings.
In each case, there is a requirement to distribute information outward, closer to the user. And in each case, there’s a parallel necessity to protect it. Corporations that are investigating the nascent software as a service (SaaS) model, for example, report that data security is their number one concern.
This is only complicated by the fact that mobile devices are able to jump from location to location, meaning that the data is always in the air, like a basketball, vulnerable to being stolen.
Hence the necessity for a strong, ubiquitous security capability at every data exchange point, from data centers to towers to ISPs, WiFis and other local service technologies.
This vulnerability isn’t new, but it will become a magnified concern as cloud computing, remote data centers and WiFi installations become the norm. And notice that we haven’t even talked about handsets yet. Irrespective of how well Intel does in the handset business, the security stack becomes a high value item. It will need to be licensed and embedded into every mobile device not because it protects the chip, but because it protects the data.
And because so much data will be flying around through the air, the security stack will require significant processing capabilities, which could compete with other functions that the processor must perform. This in turn may create a new driver for Intel’s chip business if it can demonstrate that its processors can handle the security stack with minimum impact on overall chip performance. No doubt another recent Intel acquisition, WindRiver, will be involved in this assignment.
So the net is that Intel could deliver a powerful economic argument for the continuation of Moore’s Law, based on the combination of distributed computing, tighter security, and improved integration for optimum processor performance.
This argument becomes even more compelling when you can sell it to the infrastructure providers — the telecommunications companies — who are stressed to find new ways to improve the data pipe. If the Verizons, AT&Ts and T-Mobiles of the world tell the handset makers they want to be able to ensure a secure, distributed computing environment, the handset makers will have to make that adjustment. Other big players, like Cisco, Google, Facebook and Yahoo, will be just as interested.
Intel gets its foot in the door without having to pry it open. And that’s worth many billions of dollars more than they paid.