Multicasting, the transmission of a packet to a group of hosts, is an important service for improving the efficiency and robustness of distributed systems and applications. Although multicast capability is available and widely used in local area networks, when those LANs are interconnected by storeand-forward routers, the multicast service is usually not offered across the resulting internetwork. To address this limitation, we specify extensions to two common internetwork routing algorithmsdistance-vector routing and link-state routing-to support low-delay datagram multicasting beyond a single LAN. We also describe modifications to the single-spanning-tree routing algorithm commonly used by link-layer bridges, to reduce the costs of multicasting in large extended LANs. Finally, we discuss how the use of multicast scope control and hierarchical multicast routing allows the multicast service to scale up to large internetworks.

Wireless networks are vulnerable to many identity-based attacks in which a malicious device uses forged MAC addresses to masquerade as a specific client or to create multiple illegitimate identities. For example, several link-layer services in IEEE 802.11 networks have been shown to be vulnerable to such attacks even when 802.11i/1X and other security mechanisms are deployed. In this paper we show that a transmitting device can be robustly identified by its signalprint, a tuple of signal strength values reported by access points acting as sensors. We show that, different from MAC addresses or other packet contents, attackers do not have as much control regarding the signalprints they produce. Moreover, using measurements in a testbed network, we demonstrate that signalprints are strongly correlated with the physical location of clients, with similar values found mostly in close proximity. By tagging suspicious packets with their corresponding signalprints, the network is able to robustly identify each transmitter independently of packet contents, allowing detection of a large class of identity-based attacks with high probability.

No abstract

This paper gives a method for finding a minimum spanning tree in an undirected graph. If the problem graph has n vertices and e edges, the algorithm runs in 0(e log log n) time. This time bound is the same as that of a new algorithm by Yao, but Yao's method seems more complicated to implement. A modification of the method improves the l+€ running time to 0(e), if e is fl(n) for some positive constant 6. Another algorithm finds a minimum spanning tree of a planar graph in 0(n) time. The paper also presents some results which suggest that any method for finding a minimum spanning tree requires ft(e log log n) comparisons in the worst case.

No abstract

The V distributed System was developed at Stanford University as ,part of a research project to explore issues in distributed systems. Aspects 'of the design suggest important directions for the design of future operating systems and communication systems.DAVID R. CHERITON The V distributed system is an operating system designed for a cluster of computer workstations connected by a high-performance network. The system is structured as a relatively small "distributed" kernel, a set of service modules, various run-time libraries and a set of commands, as shown in Figure 1. The kernel is distributed in that a separate copy of the kernel executes on each participating network node yet the separate copies cooperate to provide a single system abstraction of processes in address spaces communicating using a base set of communication primitives. The existence of multiple machines and network interconnection is largely transparent at the process level. The service modules implement value-added services using the basic access to hardware resources provided by the kernel. For instance, the V file server implements a UNIX-like file system using the raw disk access supported by the kernel. The various run-time libraries implement conventional language or application-to-operating system interfaces such as Pascal I/O and C stdio [Zl]. Most V applications and commands are written in terms of these conventional interfaces and are oblivious to the distributed nature of the underlying system. In fact, many programs originated in non-distributed systems and were ported with little or no modification-the original source was simply linked against the V runtime libraries. 0 1988 ACM OOl-0782/88/0300-0314 $1.50 the full power of the mainframe would have been available, such as running the simulation at night. A first tenet in our design philosophy is that highperformance communication is the most critical facility for distributed systems. By high performance, we mean providing fast exchange of significant amounts of data matching in essence the requirements of conventional file access. Slow communication facilities lead to poor performance and a proliferation of elaborate techniques for dealing with these limited facilities, analogous to the effect of slow and expensive memory on operating systems technology in the 1960s and 1970s. Fast communication allows the system to access state, such as files, without concern for location, thereby making true network transparency feasible. This is analogous to the liberating affect that low-cost memory has had on operating systems and applications since the late 1970s. Not only are the resulting systems faster, they are also simpler because there is no need to highly optimize the use of communication as a scarce resource.A second tenet of the design philosophy is that the protocols, not the software, define the system. In particular, any network node that "speaks" the system protocols (or a sensible subset) can participate, independent of its internal software architecture. Thus, the challenge is ...

The V kernel supports an abstraction of processes, with operations for interprocess communication, process management, and memory management. This abstraction is used as a software base for constructing distributed systems. As a distributed kernel, the V kernel makes intermachine boundaries largely transparent.In this environment of many cooperating processes on different machines, there are many logical groups of processes. Examples include the group of tile servers, a group of processes executing a particular job, and a group of processes executing a distributed parallel computation.In this paper we describe the extension of the V kernel to support process groups. Operations on groups include group interprocess communication, which provides an application-level abstraction of network multicast. Aspects of the implementation and performance, and initial experience with applications are discussed.

No abstract