Today, we're continuing with reader feedback to our newsletter about convergence moving to an "enterprise on a blade" in which discrete boxes such as routers, switches and PBXs are replaced by services on a blade server. One reader commented: "Enterprise on a blade is asking a lot of ASIC development. Customers have always asked vendors for a 'godbox' and to a certain degree, networking vendors have delivered." We agree that vendors have delivered, and have been doing so for years but "to a certain degree." We also think that the godbox and ASIC solutions aren't always the best choice.

On the issue of the godbox, we recall a previous discussion on the merits of an all-purpose device - a Swiss Army knife approach to networking solutions. While a Swiss Army knife can be used to cut a steak and to trim threads, it is not necessarily the best for either task. Similarly, the service access multiplexer, the enterprise blade, and the godbox may also reduce capital expenses or be more convenient to use but seldom are they the best in class to support each individual need in the enterprise.

On the issue of using ASICs wherever possible (something generally designed in an all-purpose box), our biggest concern is the possibility for rapid obsolescence. Imagine, for example, an ASIC installed in a router today that can’t support IP tagging to support IP VPN over an MPLS network, or that uses H.323 instead of SIP for VoIP. Imagine an ASIC in a gateway that can’t support presence in a unified communications application. While ASICs are well-suited for some tasks, hardware designers must also accommodate protocols that aren’t even written yet. ASICs may make blade design easier, but designers today rightly seem to be embedding as much flexibility for software upgrades that keep up with rapidly changing protocols and applications support.

Finally, the same respondent also wrote: “At some point . . . the laws of physics (politics, religion, etc.) get in the way, and the reality of a complete cloud will still never be complete. How many enterprises today are complete (or even majority) thin clients?”

We reiterate our original position on the need for premises-based solutions (with and without the enterprise blade), cloud computing, and a mix of the two -- suggesting that the best solution will vary on a business case basis in every design. We think that the well-designed business case should become the “bible” when making architectural choices.

Steve Taylor is president of Distributed Networking Associates and publisher/editor-in-chief of Webtorials. Larry Hettick is a principal analyst at Current Analysis.