Papers by Krishan sabnani
The Internet economy of the G-20 countries is estimated at roughly $4 trillion today [4]. How doe... more The Internet economy of the G-20 countries is estimated at roughly $4 trillion today [4]. How does an ecosystem that is built on top of a handful of simple protocols (such as IP, TCP, DNS, DHCP, HTTP and TLS) manage to deliver such high value? Key to this success is the openness of the Internet, based on an end-to-end design philosophy for applications and services. Any Internet application works the same way anywhere in the world, and can be used over any low-level networking technology. The network may affect performance and usability, but not functional compatibility. This emphasis on endto-end design ensures universal interoperability, and leads to an expanding (rather than fragmented) marketplace for applications and services as the Internet grows. Most importantly, applications do not have to be modified to benefit from networking improvements. This article puts forth the view that it is crucial to preserve this end-to-end design philosophy even as the tectonic forces of mobil...
IEEE Communications Magazine, 2015
Wireless Networks, 1995
This paper describes the design and performance of a link-layer protocol for indoor and outdoor w... more This paper describes the design and performance of a link-layer protocol for indoor and outdoor wireless networks. The protocol is asymmetric to reduce the processing load at the mobile, reliability is established by a combination of automatic repeat request and forward error correction, and link-layer packets are transferred appropriately during handoffs. The protocol is namedAIRMAIL (AsymmetrIc Reliable Mobile Access In Link-layer). The asymmetry is needed in the design because the mobile terminals have limited power and smaller processing capability than the base stations. The key ideas in the asymmetric protocol design consist of placing bulk of the intelligence in the base station as opposed to placing it symmetrically, in requiring the mobile terminal to combine several acknowledgments into a single acknowledgment to conserve power, and in designing the base stations to send periodic status messages, while making the acknowledgment from the mobile terminal eventdriven. The asymmetry in the protocol design results in a one-third reduction of compiled code. The forward error correction technique incorporates three levels of channel coding which interact adaptively. The motivation for using a combination of forward error correction and link-layer retransmissions is to obtain better performance in terms of end-to-end throughput and latency by correcting errors in an unreliable wireless channel in addition to end-to-end correction rather than by correcting errors only by end-to-end retransmissions. The coding overhead is changed adaptively so that bandwidth expansion due to forward error correction is minimized. Integrity of the link during handoffs (in the face of mobility) is handled by window management and state transfer. The protocol has been implemented. Experimental performance results based on the implementation are presented.
A High Speed Transport Protocol for Datagram/Virtual Circuit Networks
Computer Communication Review, 1989
An Asymmetric Protocol for Digital Cellular Communications
Describes the design, validation, implementation and performance of an asymmetric link-layer prot... more Describes the design, validation, implementation and performance of an asymmetric link-layer protocol for a wireless link. The motivation for designing a new link-layer protocol is to obtain better performance in terms of end-to-end throughput and latency by correcting errors in an unreliable wireless link in addition to end-to-end correction rather than by correcting errors only by end-to-end retransmissions. The protocol described concentrates on asymmetry, although the concept of adaptive forward error correction is briefly introduced. The protocol also supports mobility. The asymmetry is needed in the design because the wireless terminals have limited power and smaller processing capability than the base stations. The key ideas in the design consist of placing the bulk of the intelligence in the base station as opposed to placing it symmetrically, in requiring the wireless terminal to combine several acknowledgments into a single acknowledgment to conserve power, and in designing the base stations to send periodic status messages, while making the acknowledgment from the wireless terminal event driven. The asymmetry in the protocol design results in a one-third reduction of the compiled code and a two-thirds reduction of processing overhead in the wireless terminal. Some performance results are also presented
This paper presents the design and analysis of three reliable multicast transport protocols for h... more This paper presents the design and analysis of three reliable multicast transport protocols for high speed networks. The novelty of these protocols lies in the technique used in combining the acknowledgments of individual destinations along the underlying multicast tree to prevent acknowledgement implosion and in the technique used in preventing unnecessary retransmission by performing local multicasts. These protocols use the periodic exchange of complete state information NRS90] between the source and the destinations and a block-based Selective Repeat retransmission scheme to improve the overall performance in a high speed networking environment. Performance of each protocol is analyzed in terms of throughput, end-to-end delay, bu er requirement, acknowledgment tra c and retransmission tra c. Based on this analysis and the complexity of implementation, one of the three protocols is recommended for reliable multicasting in high speed networks.
IEEE Journal on Selected Areas in Communications, 1997
This paper presents the design, implementation, and performance of a reliable multicast transport... more This paper presents the design, implementation, and performance of a reliable multicast transport protocol (RMTP). RMTP is based on a hierarchical structure in which receivers are grouped into local regions or domains and in each domain there is a special receiver called a designated receiver (DR) which is responsible for sending acknowledgments periodically to the sender, for processing acknowledgment from receivers in its domain, and for retransmitting lost packets to the corresponding receivers. Since lost packets are recovered by local retransmissions as opposed to retransmissions from the original sender, end-to-end latency is significantly reduced, and the overall throughput is improved as well. Also, since only the DR's send their acknowledgments to the sender, instead of all receivers sending their acknowledgments to the sender, a single acknowledgment is generated per local region, and this prevents acknowledgment implosion. Receivers in RMTP send their acknowledgments to the DR's periodically, thereby simplifying error recovery. In addition, lost packets are recovered by selective repeat retransmissions, leading to improved throughput at the cost of minimal additional buffering at the receivers. This paper also describes the implementation of RMTP and its performance on the Internet.
Mobile Networks and Applications, 1996
In this paper, we present several challenges and innovative approaches to support nomadic computi... more In this paper, we present several challenges and innovative approaches to support nomadic computing. The nomadic computing environment is characterized by mobile users that may be connected to the network via wired or wireless means, many of whom will maintain only intermittent connectivity with the network. Furthermore, those accessing the network via wireless links will contend with limitations of the wireless media. We consider three general techniques for addressing these challenges: (1) asymmetric design of applications and protocols, (2) the use of network-based proxies which perform complex functions on behalf of mobile users, and (3) the use of pre-fetching and caching of critical data. We examine how these techniques have been applied to several systems, and present results in an attempt to quantify their relative effectiveness.
Experiences with Network-Based User Agents for Mobile Applications
Mobile Networks and Applications, 1998
Wireless networks are characterized by simple end devices and limited bandwidth. One solution to ... more Wireless networks are characterized by simple end devices and limited bandwidth. One solution to address these and other limitations of the wireless mobile environment that has been widely pursued is the placement of proxies, or agents, inside the network to assist with application processing that would normally take place on end devices. These agents can additionally manipulate data to reduce
An approach for conformance testing of protocols specified as a collection of communicating finit... more An approach for conformance testing of protocols specified as a collection of communicating finite state machines (FSMs) with two parts, pruning and a guided random walk procedure, is presented. First the protocol is pruned to various sets of machines; each set provides only one service. This significantly reduces the test sequence length. Then a guided random walk procedure that attempts to cover all transitions in the component FSMs is used. The results of applying the procedure to the full-duplex alternating bit protocol and the asynchronous transfer mode (ATM) adaptation layer convergence protocol are presented. For the ATM adaptation layer, 99% of component FSMs' edges can be covered in a test with 11692 input steps. Previous approaches cannot generate conformance tests for standard protocols (such as ATM adaptation layer) specified as a collection of communicating FSMs
IEEE Transactions on Communications, 1996
We present a new approach for conformance testing of protocols specified as a collection of commu... more We present a new approach for conformance testing of protocols specified as a collection of communicating finite state machines (FSM's). Our approach uses a guided random walk procedure. This procedure attempts to cover all transitions in the component FSM's. We also introduce the concept of observers that check some aspect of protocol behavior. We present the result of applying our method to two example protocols: full-duplex alternating bit protocol and the ATM-adaptationlayer-convergence protocol. Applying our procedure to the ATM adaptation layer, 99% of component FSM's edges can be covered in a test with 11692 input steps. Previous approaches cannot do conformance test generation for standard protocols (such as asynchronous transfer mode (ATM) adaptation layer) specified as a collection of communicating FSM's.
IEEE Journal on Selected Areas in Communications, 1997
Wireless messaging is an integral component of universal personal communication services (PCS's).... more Wireless messaging is an integral component of universal personal communication services (PCS's). Its growth is likely to be further fueled by the availability of new data capabilities in the new PCS air interfaces.
Wireless Networks, 1995
This paper describes the design and performance of a link-layer protocol for indoor and outdoor w... more This paper describes the design and performance of a link-layer protocol for indoor and outdoor wireless networks. The protocol is asymmetric to reduce the processing load at the mobile, reliability is established by a combination of automatic repeat request and forward error correction, and link-layer packets are transferred appropriately during handoffs. The protocol is named AIRMAIL (AsymmetrIc Reliable Mobile Access In Link-layer). The asymmetry is needed in the design because the mobile terminals have limited power and smaller processing capability than the base stations. The key ideas in the asymmetric protocol design consist of placing bulk of the intelligence in the base station as opposed to placing it symmetrically, in requiring the mobile terminal to combine several acknowledgments into a single acknowledgment to conserve power, and in designing the base stations to send periodic status messages, while making the acknowledgment from the mobile terminal event-driven. The asymmetry in the protocol design results in a one-third reduction of compiled code. The forward error correction technique incorporates three levels of channel coding which interact adaptively. The motivation for using a combination of forward error correction and link-layer retransmissions is to obtain better performance in terms of end-to-end throughput and latency by correcting errors in an unreliable wireless channel in addition to end-to-end correction rather than by correcting errors only by end-to-end retransmissions. The coding overhead is changed adaptively so that bandwidth expansion due to forward error correction is minimized. Integrity of the link during handoffs (in the face of mobility) is handled by window management and state transfer. The protocol has been implemented. Experimental performance results based on the implementation are presented.
Uploads
Papers by Krishan sabnani