by Michael Gentry
In 1900, mankind�s knowledge was doubling about every 34 years. By 1990, with the increase in the number of scientists, engineers and researchers, the period for doubling our knowledge was down to only six years. Now that period is only about two years. Faced with this exponential growth in knowledge, it isn�t surprising that solutions for problems can change rapidly.
In the field of information technology, let�s take a few specific examples to understand how IT solutions can be affected. The famous Moore�s Law of semiconductor technology doubles the number of transistors (or computing power) for a given price every 18 months, and that�s projected to continue for many more years into the future.
Optical transmission is in the midst of a veritable revolution. The capacity-increase rate is four times every 24 months for each optical channel or �lambda.� With dense-wave-division multiplexing, the number of lambdas is doubling (or faster) every year. Today we�re up to 160 lambdas, each carrying 10 gigabytes per second for 1.6 terabits per second in a single strand of optical fiber. Recently Nortel announced capacities of 40 gbps per lambda!
In network switching, as described in a presentation titled �Ethernet: not just for LANs anymore� (Passmore, The Burton Group Networks & Telecom Strategy Service, Sept. 18, 2000), ethernet capacity is increasing by a factor of 10 about every two to three years: fast ethernet (100 megabytes per second in 1996), gigabit ethernet in 1998 and 10-gigabit ethernet today. Based on GbE advancements, in the short span of about 12 to 18 months, a new solution (GbE) has swept the old solution (asynchronous-transfer mode) out of the campus-area-networking picture.
The point is not merely that everything in the IT world is getting �faster/better/cheaper/smaller.� The harsh truth is that actual technologies and solutions can become obsolete practically overnight. Furthermore, as technology changes at this feverish pace, requirements are evolving just as rapidly. This implies the necessity for organizational, programmatic and acquisition strategies to accommodate this pace of change.
Consider the network-centric information environment within the Defense Department called the Global Information Grid. The GIG includes all owned and leased communication and computing systems and services, software, data-security services and other associated services necessary to achieve information superiority. The GIG provides capabilities from all operating locations, including bases, posts, camps, stations, facilities, mobile platforms and deployed sites. In brief, it is DoD�s IT infrastructure.
The Army�s Enterprise Information Technology Program under the Army Knowledge Management initiative falls under the Army�s portion of the GIG. Given the pace of change we live with today, fielding �systems� into the GIG where the acquisition cycle takes five to eight years before the product is fielded is a dead strategy. User dissatisfaction and program failure is usually the result when we use this approach.
This phenomenon is clearly recognized and addressed by DoD 5000 and the Clinger-Cohen Act, which requires DoD �to the maximum extent practicable, (1) modular contracting is [to be] used, and (2) the program is [to be] implemented in phased, successive blocks, each of which meets part of the mission need and delivers a measurable benefit, independent of future blocks.� Right on! So, how can we apply this to the GIG?
First, recognize there�s no final system or solution. The GIG is evolving and will continue to evolve in capability forever, along with evolving user requirements. This isn�t spiral development. Spiral development applies to system development and implies that a finished system results from the final spiral twist. This is a continuous technology insertion and evolution of systems capability forever.
Second, adopt an organizational, funding, planning, programming, execution and acquisition strategy that supports this continuous evolution. Specifically, shift away from project managers for systems to PMs for subdomains of the GIG � for example, Army installations, Army tactical battlefield, wide-area networks, etc. Also, shift Planning, Programming, Budgeting and Execution System data from a systems orientation to a domain-sustainment and technology-insertion focus.
Finally, truly follow DoD 5000 by using a continuously evolving acquisition approach where each subdomain PM goes to industry � say, every other year � and describes the current subdomain IT situation and desired operational improvements. The subdomain PM solicits industry�s response in terms of how best to advance the subdomain capabilities during this cycle. This approach results in faster improvements and faster return on investment, and it avoids the obsolescence that comes with the systems-development and fielding approach.
The question is whether or not this is technically feasible. Will backward compatibility and interoperability problems overcome this approach? Today, this complete evolutionary approach is becoming possible as industry adopts the Internet protocol as the layer and protocol supporting:
|Technology integration (unified digital environment for computing and communications);|
|Multimedia integration (voice, data and video seamlessly handled); and|
|Standardization for a heterogeneous vendor environment to exist and work within the GIG.|
Using IP as the convergence layer is the key to preserving interoperability across the GIG over time with multiple vendors� equipment employed.
This continuous-evolution approach has precedence. It has been used and has demonstrated the ability to work well in the face of rapid change. The Army�s PM for the common-user installation-transport network, a subdomain PM vs. a system PM, was fielding ATM technology for the core of the Army�s installation network (essentially a CAN) when GbE technology arrived from industry. Within months, testing of the new GbE gear was completed, technical-architecture changes coordinated and made, and policy adjustments implemented that enabled this PM to rapidly adopt GbE in lieu of ATM. Interoperability was never a problem between the two fundamentally different Layer 2 technologies because they both carry IP packets. Users now receiving the GbE solution are exceptionally happy. Also, older ATM gear is now being quickly replaced with GbE as funds permit, and the transition is swift and smooth.
We�re in a period of rapid technological change and transition from the older circuit-switched world to a new packet-oriented world. We need organizational, programmatic, planning, budgeting and acquisition strategies to match or else we�ll constantly be fielding obsolete solutions and systems. DoD acquisition policy and regulation already supports the approach outlined here. It�s becoming technically feasible across the GIG itself and is more in tune with the realities of the 21st century�s IT environment.
As the ad says, �Life is an ongoing project,� and so is the GIG.
Dr. Gentry is the senior technical director and chief engineer at Army Signal Command, Fort Huachuca, Ariz.
Gbps � gigabytes per second
GIG � Global Information Grid
IP � Internet protocol
IT � information technology
LAN � local-area network
PM � project manager
Back issues on-line | "Most requested" articles | Article search | Subscriptions | Writer's guide
Army Communicator is part of Regimental Division, a division of Office Chief of Signal.