Originally Posted on December 01, 2005
Why an Software as a Service is not necessarily the next big thing
Before I leave the subject of abstraction from the last post, let me relate it to one other purposeful decision that NetworkStreaming has made: The Remote Support Appliance.
To explain our decision to use an appliance, though, I have to take you back to the prehistoric dawn of time, back when mainframes roamed the datacenter like herds of Triceratops. Back then, things grew big, and the IT administrator crowed that were this single, monolithic system to go down, so would the entire company’s computing capabilities. “Can we load another app on the server?” IT admin: “Ha! Of course we can!”
And then things got more complicated. And then the IT admin remembered that it might not be a good thing to rest all hopes on a single system. Redundancy, componentization, no single point of failure, distributed, load-sharing. These are just a few of the buzzwords and buzzphrases that pointed away from “one big” and toward “many small”. As applications got more and more resource intensive and specialized, it became less and less feasible to deploy them all into a single, tightly-coupled IT infrastructure.
Enter the Application Service Provider (ASP) and the excitement about software as a service, Salesfore.com, and utility computing. “Down with the do-it-yourselfer!” they cried “let us, the specialist, handle your applications for you”.
In case you missed it, the offer of ASP is the offer of abstraction. “Don’t worry about the underlying complexity and the massive datacenters and redundant Caterpillar generators that are required to run and maintain QOS for this application. Enjoy all of the benefits without any of the headache”.
Everyone seems to love this idea, especially the media. You can’t read an IT trade publication today without seeing something extolling the merits of software as a service, “the wave of the future”. Very soon, apparently, we will be buying all of our software the way we buy our electricity and sewage system.
And maybe we will, and maybe I need to join the flat earth society and try to debunk all those forged man on the moon pictures. I’m joking, but half serious. Technology trends are devilishly hard to pin.
I and NetworkStreaming are betting pretty heavily on software as a service not being the next big thing. The ASP model is not a bad idea; in fact it is a very good method for abstracting complexity away from the end user. However, there are several fundamental problems with the model that all the kings horses and all the kings men have not been able to resolve.
I’ll pick up where I’m leaving this off in the next post, but for now, ask yourself a few questions:
- What are the names of the hottest new software company and the hottest old software company? (hints: rhymes with ‘oogle’. opposite of big hard.)
- Which deployment methods are these two, hot software companies known for? Traditional server/desktop software or software as a service? (hint: one for each)
- When deploying an enterprise application, what other delivery models have these two companies chosen? (hint: appliance. appliance.)
Stay tuned for Part II of More and More Abstract (Why an Appliance for Remote Support)
