Tag Archives: hardware

I know it’s a big debate right now, but centralized network computing will win in the end.

Centralized network computing is the term used to describe a system of networked web servers (or a single web server) that provides integrated applications and storage for multiple users who can access the system through distributed terminals. The system can be geographically distributed or not, but will share a common (integrated) network of applications, probably using a software interface standard to encourage and enable multiple independent application development teams.

Centralized networks are inevitable because of self-reinforcing competitive advantages. Economies of scale and market forces will lead to substantial change in the way we compute, and the systems we use now are simply the seeds that will grow into (and merge together to form) global centralized information service providers. There are already some very strong indicators that this trend is happening, and the potential points to this trend being a very long one.

1. There are economies of scale in processing Load balancing can optimize processor utilization and provide both faster interactivity and reduced hardware investment.
2. There are competitive advantages in information and application aggregation Integrations can break down the walls between programs, improving functionality through better integration and data sharing. You can analyze data in more dimensions and with more flexibility. Development rates can improve as it becomes possible to support separation of more software components and the people that work on them.
3. Load balancing improves transmissions Transfer rates improve because fewer nodes are required and data traffic can be optimized. Information served to you from a server on your edge is more reliable and fast than information sent to you from another server through your edge.
4. End-user transparency The front-end possibilities under centralized computing are not limited beyond that of other systems. This implies that there will not be a selection bias away from centralized systems because end-users will not prefer or recognize the difference between systems. That is not to say that they will all be the same – only that they all could be. The opportunity set in one system is available in the other.
5. The outsourced storage industry exists This implies that there is a willingness to adopt on the part of the owners of data.

You can see the markets already rewarding companies that are moving to take advantage of this trend. Many of these companies are providing application services along with ISP connectivity, and they are capturing traffic. This traffic is investing time and thought into signaling their own preferences. Some examples include personalizing page layout and content — often even using system-wide wallets and e-mail. Giving users what they prefer is a huge competitive advantage. The time it takes to personalize a competing system is a high transaction cost – especially relative to the low cost of inertia.

Eventually, you will be using only a browser. All your computing will occur on a centralized system and be sent to your browser for translation into your interface. All of your applications will be centrally hosted, so your profile and applications – essentially everything you do on your computer – will be available to you from any machine, at any time.

Multiple systems will compete for scale, reducing marginal costs and creating meaningful and irreproducible competitive advantages. This race will likely be won permanently by 2050. Before that time, ASP services will consolidate rapidly along with economic cycles. the early players will rely on loss leader models to attract user bases, and will transition to profitability as the scale reaches the tipping point. The companies that make it to profitability first and reinvest in their technology platform will improve their integration, breadth, and quality to further support their competitive advantages.

In the first decade or two of this trend, there will probably be dozens of smaller companies that are able to enter and gain market share against their larger competitors. These companies will have competitive advantages most likely based on data storage, traditional media integration, wireless adoption, software platform architecture, applications suite integrations, and possibly international comparative advantage. After 20 years, the marginal advantages possible from these characteristics will not pose a meaningful threat to the aggregation and scale advantages of the top few market participants.

Consolidate or die will be the mantra of information companies.