Consideration for the “big picture” is a theme that runs deep among all of the Military’s advanced programs. The push is on to leverage technologies, common computing standards and design concepts across multiple platforms. The Navy’s DD(X) program exemplifies that trend. The Navy’s new DD(X) program is the centerpiece in a family of three surface combatant ships, including a destroyer (Figure 1), a cruiser and a smaller craft for littoral operations. Over the life of the program, the DD(X) contract is expected to total $100 billion for construction of around 70 warships comprising the DD(X) family. The cruiser and destroyer are expected to share a common hull design.

At the heart of the DD(X) is the concept of a Navy standard combat system. That system is to be used first on DD(X), and then on other families of future ships. The system could then potentially also be back-fit onto existing ships now in service. A central element of the standard combat system is an onboard computing network that requires a substantially smaller number of IT personnel to the run and administer it.

To meet that goal, the DD(X) developers came up with the idea of a modular, pre-integrated server core. That core is essentially a room full of server racks the size of the back of a van. Rugged computers aren’t required in this “room,” because all the ruggedness—the shock isolation, the cooling, the heat and power are all handled at the room level. Filled with around 15 or so racks of server-level computers, this core server room is dropped right into the ship during construction and later upgrade phases.

That strategy makes for a more cost-effective and schedule-friendly way of integrating computing elements on a ship. The approach is a dramatic improvement over the current norm for shipboard computing gear. Today’s ships are content with a rat’s nest of cables going between rooms and all myriad of stand-alone computing stations. In contrast, the DD(X) server core links over a small fiber-optic cable going throughout the ship. It’s much lighter, easier to maintain and more survivable than what’s aboard today’s Navy vessels.

Layered Architecture

With the server core at the center, the DD(X) network employs a layered architecture where more of the applications are run in that server core with a high-speed network linked to the various nodes throughout the ship that interface with sensors, weapons and so on. Using a services-oriented architecture, the layered scheme isolates, through middleware, the operating system from the services. That isolation makes is easy to change out the processors with only minor code changes.

The initial core servers for DD(X) were Sun Microsystems SPARC servers running Solaris. The DD(X) program’s business model calls for a standards-based competition plan that competes every four years, the most recent of which was concluded late last year. After narrowing the field down to EMC, Hewlett-Packard, Sun and IBM, the final winner was IBM’s Opeteron Blade Servers with Red Hat Linux. Compared to its competitors, IBM reportedly had the best cost-model and the best “here and now” high-performance processor solutions.

For the local embedded computing for the ships sensors and weapons electronics, the DD(X) program favors Compact PCI single board computers. The software team has compiled a list of those SBCs in a Total Ship Computing Environment catalog. With LynxOS as their RTOS, those SBCs run the same middleware as DD(X) core servers. The goal is to move as much processing as possible into the core servers, and leave only the barest functionality on the embedded SBCs.

Raytheon Leading the Software Team

Responsible for leading the massive software development work for the DD(X) program is Raytheon Integrated Defense Systems. According to Bob Martin, Director of System Software Development for DD(X), the program represents the biggest software project that Raytheon has ever done. Even now in the relatively early stages of the program, the company has over a thousand software engineers on the job, including team members from most all of the major defense contractors and a host of smaller companies.

Part of Raytheon’s role is to work on the software process with some of the small vendors. For its part, Raytheon is a CMMI Level 4-rated software developer. CMMI (Carnegie-Mellon Software Maturity) is a level of quality and predictability, with ranks ranging from 1 to 5 (5 being the highest). On the DD(X) program, Raytheon is bringing in software developers from small vendors in the ship segment to train and mentor them to get them up to at least Level 3. That’s key because for small companies that focus on marine-navigation radar, ship steering control systems and so forth, the world of complex leading-edge software is brand new.

Four years into its work, the DD(X) software development is actually on cost and on schedule—a rare feat in the software realm. And while the DD(X) did face some cuts in the budgets earlier this year, the software portion emerged unscathed. Part of that is because it has performed so well, but that’s also indicative of the importance of the computing/software environment: it’s what provides the automation in the application layer that allows for smaller IT staff aboard ship, and the reduction of costs associated with that.

Spiraling Out Pieces of DD(X)

Like many of the Military’s major programs, DD(X) uses a spiral development strategy that enables pieces of its functionality to get implemented on an incremental basis. Pieces of the TSCEi (Total Ship Computing Environment infrastructure) used in DD(X), which comprises the hardware, software middleware—can be leveraged on board existing vessels. Some of the Navy’s amphibious ships are deploying a new Ships Self-Defense System (SSDS) in 2008, and pieces of the TSCEi will be used in that. Pieces of the TSCEi have also been given to the Aegis Destroyer people, also for deployment in 2008.

The idea is cost savings so the Navy doesn’t have to invent a new infrastructure of computing environment for every new ship class.

Java is a key technology for the DD(X). The Navy’s Open Architecture Computing Environment (NOACE) has been made the standard for all future software systems on Navy warships. That includes shipboard weapon systems, such as anti-aircraft cannon controls as well as avionics systems aboard naval aircraft. The standard calls for all new software to develop in either C++ or Java, and makes specific mention of moving away from Ada. In DD(X) software developers have pushed the envelope of Java real-time garbage collection technology. Release 4 of the DD(X) software environment will switch completely to a real-time Java VM.

Zumwalt the First DD(X) Ship

In early April of this year, the Navy announced that the first DD(X) destroyer will be designated DDG 1000. As the lead ship in the class, it will also be named “Zumwalt” in honor of former Chief of Naval Operations (CNO) Admiral Elmo R. Zumwalt, Jr. Developed under the DD(X) destroyer program, Zumwalt is the lead ship in a class of next-generation, multi-mission surface combatants tailored for land attack and littoral dominance.

Compared to current U.S. Navy destroyers, the Zumwalt class destroyer will triple both current naval surface fires coverage as well as capability against anti-ship cruise missiles. It has a 50-fold radar cross section reduction from current destroyers, improves strike group defense tenfold and has 10 times the operating area in shallow water regions against mines.

Last year, Congress fully supported the DD(X) budget request, and the Zumwalt class is ready to start construction. In November 2005, the Department of Defense granted Milestone B approval, authorizing entrance into Phase IV of the program, including the detail design and construction of the two lead ships. Under the Navy’s dual lead ship acquisition strategy proposed in the President’s budget for fiscal year 2007, Northrop Grumman Ship Systems and General Dynamics Bath Iron Works will concurrently build the dual lead ships. Zumwalt is scheduled to be delivered in 2012.