In the early years of computing, real-time systems consisted of single computers interfaced to a few sensors and effectors, and perhaps to a user control device as well. While such systems still exist, and have become almost ubiquitous in a myriad of everyday devices, the world of real-time computing has expanded in scope and complexity in ways comparable to the proliferation of computers for many other aspects of business, economic and social life. Perhaps the most recent evidence of this expansion is the employment of distributed processing to achieve the objectives of real-time system designers. With the expanded use of computers, particularly in distributed real-time embedded systems (DREs), issues of complexity have emerged that were not faced by early real-time system designers. Ccomplex aggregations of computers operating as a distributed system have assumed a distinctly different character from early real-time systems, exhibiting far more variation in form as well as in the underlying technology base. Dealing with this complexity has become a major driver for innovation in both system architecture and computing technology. The challenge designers face today is, how to design, build, deploy and maintain ever larger and more capable distributed real-time embedded systems to meet the demands of a changing world. Modern real-time systems are likely to be more than just systems; they are rapidly becoming large-scale systems of systems. Because of their long life and high cost, complex real-time systems cannot be recreated as new starts when new requirements emerge. They must evolve and undergo changes and upgrades in place. Furthermore, since the technology base on which DREs are built is changing rapidly, complex real-time systems must allow for the convenient insertion of new computing equipment and supporting software products. For these reasons, flexibility of design has become a paramount concern. This paper will present major challenges and promising solutions for the problem of building real-time open architectures. Based on our experience within the Navy shipboard computing systems domain, we will consider technical characteristics, change management, system flexibility and dynamic resource management.