Cloud Computing Commentary

Please comment on Cloud Computing with GPU Intensive Apps (current state and future viability).

Admittedly, this topic isn’t my long suit. Cloud Computing is kind of the übercategory for delivery of IT capabilities over the Internet. Most of the business offerings in that category aren’t computationally intensive, particularly for graphics.

When thinking Internet-delivered capabilities, I’m presuming some sort of thin device on the client-side. That doesn’t generally mesh well with a heavy-duty GPU and the corresponding bandwidth requirements to feed it. Is this for Internet delivery of multimedia gaming or other real-time graphics applications?

Perhaps I’m just a troglodyte, but I don’t see this in the present Cloud Computing environment and would be somewhat apprehensive about opining a future without knowing more specifics. Reader comments are most welcome.

SaaS Commentary

Please comment on Software as a Service (current state and future viability).

At the risk of placing myself firmly in the pantheon of IT curmudgeons, I think that SaaS is another new name for an old concept. In 1993, we referred to this sort of thing as an outsourced or hosted application (admittedly not particularly sexy names).  IT lexicon purists may split hairs over the use of a common code base, but fundamentally it’s the same thing.

In my view, SaaS is simply another way to add a particular business capability. There are certain conditions where it makes sense to adopt this approach. For highly differentiated capabilities, it probably isn’t a good option. For more standardized capabilities, it is. Lower investment and speed of implementation are also generally considered as advantages for SaaS solutions.

There are other considerations as well that would be built into a scorecard that would be used to rate the suitability of a SaaS solution versus other options. Again, it’s all about what you are trying to accomplish and what that means about how you should do it.

Obviously, SaaS is and has been a viable option for provisioning capabilities. In the future, my opinion is that it will most likely be subsumed into the larger utility or cloud computing category.

SOA Commentary

Please comment on Services-Oriented Architecture (current state and future viability).

Services is one of the most overused and least commonly defined words in IT today (discussing Services in both SOA and Managed Services contexts at the same time makes for a challenging conversation). A close runner-up is Architecture.

Having them both in the name of an IT construct still somewhat frightens me.

My comments today on SOA will be brief. I believe that SOA is a new, fancy IT-ism for an older concept called Enterprise Architecture carried to its logical conclusion or goal.

Enterprise Architecture is all about common business services as instantiated in technology. That happens via an end-to-end EA process reaching back to the business. SOA simply articulates that differently. The technology behind them is the same.

I’m not sure if SOA has the staying power of EA. If pressed, I’d speculate that SOA would eventually diminish in the IT lexicon, as EA efforts are successful in reaching the same goal using processes in place now.