Is the 'god box' finally real?
Once a skeptic, now a believer, Jon Oltsik says the single-system concept is closer to reality than ever before.
The term "god box" has its own definition in the technology industry. For those unfamiliar with the phrase, a "god box" is a single system packed with ridiculous amounts of functionality. Think of an all-in-one television, cable box, TiVo, and home entertainment speaker system and you get the idea. One "god box" always replaces numerous more pedestrian systems.
In the past, it was not unusual for some new company to come to me--an analyst--and spin its new proprietary "god box." But I always remained skeptical. Why? I assumed that a "god box" would fall victim to either technology or economic limitations. That is to say, "god boxes" were always either underpowered or overpriced. Additionally, you could usually get better software functionality from specialists; "god boxes" were more of a software aggregation than a quality play.
Times have changed. No, we aren't quite yet at a point where the industry can produce Scott McNealy's visionary, "big honkin' Web switch," but we are getting closer. "God boxes" are more real than ever because of:
Multicore processors. In the past, there was no getting around the need for multiple expensive physical processors. As a result, phat Hewlett-Packard, IBM, and Sun Microsystems servers were kings of the system hill, but they were usually reserved for back-office transaction processing or high-performance computing. Enter commodity multicore processors and, voila, big horsepower/low price. I can now do fancy processing algorithms on a modestly priced box.
Hardware commodification. Hardware essentials such as memory, Ethernet ports, FPGAs (field-programmable gate arrays), or fixed-function processors are all readily available for a few hundred bucks at most. To continue to compete, hardware widget makers constantly increase the functionality of their own wares, thus accelerating the entire intelligent system (aka "god box") cycle.
Software critical mass. Four things have happened on the software front: 1. Design software quality is up, making entry-level programs "good enough" for lots of tasks. 2. Open source frees developers to outsource software "plumbing" and focus on value-added features. 3. Component software speeds incremental innovation. 4. The army of global programming talent available for hire today is much larger than it was. Server virtualization from Citrix, Microsoft, and VMware will also fuel this fire.
These advances change the playing field and make the "god box" concept increasingly real. Small start-up hardware jocks like A10 Networks deliver lightning fast multifunction systems. Cisco Systems keeps adding new blade-based functionality to its Catalyst Ethernet switches, Juniper continues to enhance its JUNOS operating system with advanced packet processing, and SonicWall can deliver an all-in-one networking/security appliance in a small form factor. These are just a few examples.
Over the next few years, what we know today as routers, switches, and Unified Threat Management (UTM) appliances may all become multifunctional "god boxes" that combine the best operating systems with advanced hardware to deliver integrated functionality that boggles the mind. This will take place at the high and low end of the market. What's more, I expect to see these "god boxes" start to include industry-specific functionality. For example, networking equipment for manufacturers and logistics companies will ship with integrated radio frequency identification readers, XML rule sets, and integration into production databases.
It is worth noting that "god boxes" really favor the networking crowd. The more intelligence they can pack into the network, the less smarts you need elsewhere.
I was a skeptic, now I'm a believer. Welcome to the "god box" era. Expect to see some really big systems and really cool stuff soon.