EHR Cloud Computing Meets Moore's Law
EHR Cloud Computing Meets Moore's Law
In an increasingly computerized world, understanding why and how computers do what they do can have a direct bearing on how you evaluate the effects on care that are the result of having the computer in the loop. Also, in the future, today's medical students are going to be making decisions about computerization and other technology. What will form the basis for their decisions — their "knowledge" of Facebook or how to text with two fingers? This raises the question: How knowledgeable, not "computer-literate," but knowledgeable are you?
Take cloud computing for example. It's the latest high-tech buzzword but do you have any idea what that is? A new national survey by Wakefield Research, commissioned by Citrix, suggests that the answer is "not really" — over half of respondents, including a majority of [supposedly computer literate] Millennials, believe that stormy weather can interfere with cloud computing.
How about Moore's Law? In 1965, Gordon Moore observed that technology seemed to be enabling the number of transistors on a microchip to double every two years. In 1970, Carver Meade dubbed this "Moore's Law." A graph of microchip production suggests that Moore was correct. Some physicists believe that Moore's Law, while it describes the past, does not predict the future. As transistors are made ever smaller they approach the size at which atomic forces come into play, effectively defining a minimum size, and hence a maximum speed, of the circuits in question. This limit may be reached in the next five years to 12 years.
Even if Moore's law should fail, faster computers are not precluded, but speed will have to be achieved in some other way such as deploying multiple processors working in parallel. I will return to Moore's Law in a moment, but first a discussion about the cloud.
The cloud is double talk — just a shorthand for computing that takes place somewhere other than inside your computer whether that be actual computations, data storage and retrieval, communications, or search. Some bright folks thought that the concept of "elsewhere" was too nebulous so. Instead they chose a term that connotes that most nebulous of things — a cloud.
The aspect of the cloud that is important to you, especially if you are considering a cloud-based EHR, is that the cloud is at the other end of a narrow pipe. Every bit of information that you send or request has got to pass through that pipe. While it is easy and cheap to buy a faster computer, it is neither easy nor cheap to buy a faster pipe. The speed and capacity of the pipe has not been doubling every two years, it has been doubling every eight years to 10 years and there is no guarantee when or if it will double again. Conclusion, connection speeds never have, and never will, experience the exponential growth of chips.
When you are using remote computing resources, the speed of your local computer is almost irrelevant — well not completely, but if it is a recent model it is. Speed is dependent on two factors: 1) how much processing capacity the remote site is willing to allocate to YOU; and 2) the bandwidth (speed) of your connection.
There is a third piece to this story — application complexity. Every time you ask your EHR vendor for a new feature or the feds add certification criteria that require additional program code, the demands placed on the system increase. If the code is running locally, you may be able to offset the increased demand with a faster processor.
If the code is running in the cloud and if it requires more data to be transferred back and forth between your location and the vendor’s site, then the performance that you experience will get progressively worse as the applications get more complex.
Since this slowdown has to do with the speed of the connection, not the speed of my computer, Moore's Law is irrelevant. If all computing is going to be cloud computing, I will never again need to buy a new computer — and I will become increasingly dissatisfied as the speed continues to degrade.
This is one of many reasons that I choose, as much as possible, to keep my computing local. You might want to carefully consider any decision to outsource your EHR to the cloud. Assuming that you choose an EHR that has the potential to "speed you up" if run locally, it could slow you down if run remotely, in the cloud.
Find out more about Dan Essin and our other Practice Notes bloggers.