Pitro Zafiropulo
IBM
Network
Latest external collaboration on country level. Dive into details by clicking on the dots.
Publication
Featured researches published by Pitro Zafiropulo.
international conference on computer communications | 1990
Hans R. Muller; Mehdi Nassehi; Johnny W. Wong; Erwin Zurfluh; Werner Bux; Pitro Zafiropulo
Distributed-queue multiple-access (DQMA) and cyclic-reservation multiple-access (CRMA) access schemes for gigabit-per-second local and metropolitan area networks are described. These schemes are based on a slotted unidirectional bus structure, with both folded and dual-bus configurations. Like the distributed-queue dual-bus (DQDB) scheme, both the DQMA and the CRMA access schemes achieve full throughput independent of network speed and distance. They have two significant advantages over DQDB. They provide throughput fairness even at high speeds and large distances, where DQDB exhibits dramatic unfairness, and, by allowing reservation of multiple consecutive slots, they make segment labeling unnecessary and facilitate packet reassembly significantly.<<ETX>>
IEEE Transactions on Communications | 1978
Pitro Zafiropulo
A protocol is a set of rules which governs the interaction between processes. It is difficult to design protocols without errors because there are usually more interactions possible than anticipated. We are concerned with two-process protocols. A technique is described which identifies design errors in protocols. The technique is based on modeling a protocol as a pair of interacting graphs whereby interaction sequences are represented by path pairs (one path in each graph). The technique is currently limited to protocols that must revert to an initial or quiescent state after a finite number of interaction steps. The work represents a theory that can be automated on a digital computer.
international zurich seminar on digital communications | 1990
H.R. van As; J.W. Wong; Pitro Zafiropulo
In the spring of 1989, the IEEE 802.6 Metropolitan Area Network Standards Committee chose to modify the queue-arbitrated DQDB (Distributed Queue Dual Bus) protocol. The heavy load and overload behavior of this DQDB version is examined. Extensive simulation studies show the protocol behavior to be extremely complex so that this study is restricted to a carefully selected set of scenarios that present dominating behavioral features of the protocol. It is shown that the protocol is unfair, that node throughput is unpredictable during heavy load, and that the priority mechanism cannot be relied upon. It is also shown that network configurations exist which exhibit inverse priority behavior, i.e. less throughput with high priority than with low priority. A comparison of DQDB and reservationless slotted bus operation is presented. It is concluded that the modified DQDB media access protocol exhibits serious deficiencies during heavy load and overload conditions and that at loads low enough to reduce the damage of these adversities sufficiently the queue arbitration shows no advantage over simpler protocols such as slotted bus without reservation.<<ETX>>
IEEE Transactions on Communications | 1974
Pitro Zafiropulo
One of the basic terminal-attachment configurations is the loop. As such, it is receiving substantial attention. But it has been pointed out that loop communication networks are relatively vulnerable and unreliable. This paper considers the single-loop reliability problem. Three loop reliability (availability) improvement techniques are investigated in detail. All three techniques are based on using an auxilliary transmission path parallel to the main loop. A comparative study shows the best techniques.
Computer Networks | 1978
Harry Rudin; Colin H. West; Pitro Zafiropulo
Abstract The progress of one research effort in computer-automated protocol validation is summarized in this paper. It brings together in one place the results of a number of studies, some published and some to be published but all widely scattered in the literature. Described are a formal representation, the “duologue”, “phase-diagram”, and “state-pertubation” methods. These are discussed on the basis of a common example. Sample results from the validation of the CCITT X.21 interface are explained as an indication of what these techniques can achieve.
Computer Networks and Isdn Systems | 1994
H.R. van As; Wolfram Lemppenau; Hans Rudolf Schindler; Pitro Zafiropulo
Abstract CRMA-II is a MAC protocol for Gb/s LANs and MANs that are based on a single or dual-ring, with bus structures being covered as well. Data are transmitted in slots and buffer-insertion techniques are applied whenever data frames must be transmitted in contiguous slots or when immediate access to the media is required. Spatial slot reuse is achieved by freeing slots at their destination. Access to the media is controlled by a simple busy/free mechanism; it is enhanced by a reservation-based fairness mechanism. This results in fast access when the network is lightly loaded, while fair access and a tight delay bound are achieved during periods of sustained heavy load or overload. Fairness scheduling automatically adapts to the traffic requested by the nodes and causes minimal control intervention. Following current trends in communications, the standard ATM cell format has been adopted for both the medium and the attachment interfaces. CRMA-II control is only visible in an access sublayer between the standard ATM layer and the ATM transmission technology.
Archive | 1982
Pitro Zafiropulo; Colin H. West; Harry Rudin; D. D. Cowan; Daniel Brand
The growing trends both to increase the sophistication of functions implemented in information-handling systems and to distribute these functions in different processes has resulted in an enormous growth in complexity. This complexity is particularly acute in the interactions or protocols which specify how these processes are synchronized and communicate with one another. However, formal methods are gradually being introduced to describe these interactions; see chapters 20, 21, 23, and 25 as well as [l]–[5].
IEEE Transactions on Communications | 1973
Pitro Zafiropulo
The rapidly growing interest in loop communication systems is accompanied by an increasing need to improve their reliability and reduce their vulnerability. This paper analyzes the reliability of multistage loop networks. Analytical expressions for topologies exhibiting maximum availability are obtained in terms of key-component down-time ratios. These optimal topologies are expressed in terms of terminal connections per subloop and subloops per higher level loops. The availability performance of these optimal topologies is investigated. Finally, the availability improvement obtainable by going to higher stage loop networks is investigated.
computer software and applications conference | 1978
Pitro Zafiropulo
This paper addresses the problem of protocol failures that are caused by logically incompletely specified designs. A design is logically complete if it specifies all occurrable event (message) receptions under normal operating conditions. A more stringent requirement is that of dual completeness. A design is dually complete if it neither contains unspecified event receptions nor unspecified event transmissions. Rules for creating dually-complete designs are established and a design methodology based on these rules is proposed. The problem of designing interactions that involve interrupt-masked computations is also treated.
Archive | 1991
Hans R. Muller; Mehdi Nassehi; Johnny W. Wong; Erwin Zurfluh; Werner Bux; Pitro Zafiropulo
Distributed-Queue Multiple-Access (DQMA) and Cyclic-Reservation Multiple-Access (CRMA) are new access schemes for Gbit/s local and metropolitan area networks. These schemes are based on slotted unidirectional bus structure, both folded and dual-bus configurations. Like the DQDB, both DQMA and CRMA access schemes achieve full throughput independent of network speed and distance. However, they have two significant advantages over DQDB. First, they provide throughput fairness even at high speeds and large distances. Secondly, by allowing reservation of multiple consecutive slots, they make segment labeling unnecessary, and facilitate packet reassembly significantly.