Packet Switching in the MICRONET Network Computer

Packet switching is a communication technology which has been used extensively in geographically distributed computer networks. It is also applicable to the communication subnetworks of compact multimicrocomputers known as network computers. This paper describes the use of the language Concurrent Pascal to build a packet-switching subsystem for the MICRONET network of DEC LSI-11 microcomputers. Examples of actual Concurrent Pascal source code taken from the system demonstrate the usefulness of high-level languages with abstract data types for complex communication software.     

Multicast Communication on Network Computers

Channel-oriented packet casting is a predominant feature of Micros, an operating system designed to explore control and communication techniques for network computers containing thousands of hosts.     

Operating Systems for the Micronet Network Computer

The Micros operating system executes on a modular multimicroprocessor system. It provides system-wide high-level control as well as local operating systems for individual nodes.     

Type‐safe concurrent resource sharing

Concurrent systems often have many processes sharing a common set of resources, both memory regions and hardware devices. Among the many challenges in producing safe concurrent software are single access, atomic transactions, starvation, and deadlock. Locks are frequently used to provide single access to shared resources, but do not guarantee safe usage. This paper extends the previous work on linear, singleton, and arithmetic types and linear memory primitives. Our contributions are capabilities for shared resources, and locks to control these capabilities in provably safe ways. We present formalized locks in a lambda calculus along with the soundness properties of preservation and progress. The type system described here prevents data races. The formalized locks have also been implemented in a C‐like language and used in a network device driver. Copyright © 2010 John Wiley & Sons, Ltd.     

On the implications of routing metric staleness in delay tolerant networks

Delay Tolerant Network (DTN) routing addresses challenges of providing end-to-end service where end-to-end data forwarding paths may not exist. The performance of current DTN routing protocols is often limited by routing metric “staleness”, i.e., routing information that becomes out-of-date or inaccurate because of long propagation delays. Our previous work, ParaNets, proposed a new opportunistic network architecture in which the data channel is augmented by a thin end-to-end control channel. The control channel is adequate for the exchange of control traffic, but not data. In this paper we present Cloud Routing, a routing solution for the ParaNets architecture. We motivate the need for such a solution, not only because of stale routing metrics, but also because of congestion that can occur in DTNs. Unable to use up-to-date routing metrics to limit congestion, existing DTN routing solutions suffer from low goodput and long data delivery delays. We show how Cloud Routing avoids congestion by smart use of forwarding opportunities based on up-to-date routing metrics. We evaluate our solution using extensive OPNET simulations. Cloud Routing extends network performance past what is currently possible and motivates a new class of globally cognizant DTN routing solutions.     

Computer networks and distributed systems

Three major justifications for distributed computing-sharing physically distributed resources, combining computers for fast solutions, and providing reliability through replication-are discussed. Distributed computing milestones from 1969 to 1991 are examined, focusing on the ARPAnet national research network, Ethernet and token-ring local area networks, and workstation networks united by distributed systems software. Three themes that dominate current trends in distributed systems and computer networks are examined. They comprise tapping the immense data-carrying potential of optical fibers, efficiently using tightly coupled networks of thousands of computers, and making network access inexpensive so many people will buy services. Developments for the next decade are predicted by extrapolating from these trends