From eyh5@ee.cornell.edu Wed Nov 7 14:02:58 2001 Return-Path: Received: from memphis.ece.cornell.edu (memphis.ece.cornell.edu [128.84.81.8]) by sundial.cs.cornell.edu (8.11.3/8.11.3/M-3.7) with ESMTP id fA7J2uR01897 for ; Wed, 7 Nov 2001 14:02:56 -0500 (EST) Received: from photon.ece.cornell.edu (photon.ece.cornell.edu [128.84.81.138]) by memphis.ece.cornell.edu (8.11.6/8.11.2) with ESMTP id fA7J23J06556 for ; Wed, 7 Nov 2001 14:02:03 -0500 Date: Wed, 7 Nov 2001 14:02:15 -0500 (EST) From: Edward Hua X-X-Sender: To: Subject: 615 Paper # 38 Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Adaptive Protocols for Information Dissemination in Wirless Sensor Networks Wendi Rabiner Heinzelman, Joanna Kulik, and Hari Balakrishnan The paper proposes a family of protocols to disseminate data in a wireless sensor network that places emphasis on the communication throughput, overhead, and power consumption. SPIN, or Sensor Protocols for Information via Negotiation, aims to address the three critical issues in any wireless ad hoc network: implosion of data packets, overlapping of coverage, an resource scarcity. To overcome these inefficiencies, SPIN devises two strategies: negotiation (based on the so-called meta-data) and resource adaptation. Negotiation is done through the use of meta-data to ensure that only information that other nodes do not yet possess will be disseminated from the node, thereby reducing the amount of replicate data traversing the network. Resource adaptation is implemented in SPIN-II, which, based on a pre-set low-power threshold, determines intelligently whether the node has the sufficient resources to participate in data exchange with other nodes in the network. SPIN generates three types of messages: new data advertisement (ADV), request for data (REQ), and data message (DATA). An ADV is sent from one node (X)to its neighbors whenever it has gathered new data. It advertises to them its posession of new data and asks whether the neighbors would like to have a copy of it as well. If a neighbor is interested in getting a copy, it sends a REQ back to Node X to request the new data. X then sends the DATA with the new data to the requesting node, which, after receipt and incorporation of this new data to its own database, may launch an ADV to its neighbors (excluding X) of this new knowledge, and the process repeats itself. By repeatedly doing so, it is ensured that every node in the sensor network will, at least for a certain period of time, will possess the same knowledge as any other node of the network. This tactic has the advantage of lowering the replication of data traversing the network and thus reducing the potentially high power consumption and shortening the latency. The benefits are well demonstrated in the simulation evaluations. However, whether this is a practical solution is open for debates. Every node possessing all the accumulated knowledge of the entire sensor network, however noble, may be an overkill in implementation. A more realistic approach is to have each of the nodes possessing the knowledge of its coverage area. The ADV-REQ-DATA scheme may be invoked when some critical information, to be pre-determined before the deployment of the network, is captured and needs to be made aware by all members of the network. This may prodcue a more efficient utilization in terms of power and bandwidth consumption. Another observation is the fate of those nodes in SPIN II that have go below the preset low-energy threshold. What happens to those nodes? Do they have battery-recharging capability or are they just let die? The authors did not mention what would happen to these nodes. According to the design, these nodes will not be able to handle any DATA messages, either incoming or outgoing, given their low power level. If these nodes are indeed let die (as it is unlikely that they are equipped a battery recharger), that will introduce a number of problems to the overall performance of the sensor network. Once shortcoming is the partitioning of the network, in which nodes on two sides of the network become disconnected and can no long share information with each other. How to address the dying out issue will have a great impact on the feasibility of this scheme for commercial deployment. From andre@CS.Cornell.EDU Wed Nov 7 18:12:53 2001 Return-Path: Received: from postoffice.mail.cornell.edu (postoffice.mail.cornell.edu [132.236.56.7]) by sundial.cs.cornell.edu (8.11.3/8.11.3/M-3.7) with ESMTP id fA7NCqR01997; Wed, 7 Nov 2001 18:12:52 -0500 (EST) Received: from khaffy (d2095.dialup.cornell.edu [132.236.155.95]) by postoffice.mail.cornell.edu (8.9.3/8.9.3) with ESMTP id SAA29148; Wed, 7 Nov 2001 18:12:49 -0500 (EST) Received: from andre by khaffy with local (Exim 3.31 #1 (Debian)) id 161WHo-0000P0-00; Wed, 07 Nov 2001 18:14:48 +0100 Date: Wed, 7 Nov 2001 18:14:48 +0100 From: =?iso-8859-1?Q?Andr=E9?= Allavena To: egs@CS.Cornell.EDU Cc: andre@CS.Cornell.EDU Subject: 615 PAPER 38 Message-ID: <20011107181448.A1518@khaffy> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Disposition: inline Content-Transfer-Encoding: 8bit User-Agent: Mutt/1.3.20i Sender: =?iso-8859-1?Q?Andr=E9_Allavena?= Adaptive Protocols for INformation Dissemination in Wireless Sensor Networks This paper presents one (two) protocols for disseminating data in a wireless sensor network. The goal is to save energy by avoiding duplicate transmissions. Their model assume a fixed cost for each peer to peer communication (no broadast, cost to send from A to B1, B2, .. Bn will be n, not 1). The way they do so is to have each node which has data to share send an advertisement describing his data. Upon request by some of its neighbours, it will send them only the needed parts. Since the meta data (the description of its data is smuch smaller than the data itself, this is much better). Of course their scheme is much more (energy) efficient than gossiping or floding. It is reliable (data will be disseminated even if some nodes fail, provided the network is partioned). Their scheme is really closed to an inverted RTS/CTS (request to send packets x1, x2,..., xn), no anwser means don't send anything or yes send me something. I think we have seen somewhere a routing protocol doing the same. I think most of the wireless sensor network can broadcast, in which their results don't hold anymore (a classic floding is much cheaper than wat they say). Further more, I don't really understand the point of disseminating data in the network. Aggregating, yes. -- André Allavena (local) 154 A Valentine Place École Centrale Paris (France) Ithaca NY 14850 USA Cornell University (NY) (permanent) 879 Route de Beausoleil PhD in Computer Science 06320 La Turbie FRANCE From ramasv@CS.Cornell.EDU Wed Nov 7 22:35:57 2001 Return-Path: Received: from exchange.cs.cornell.edu (exchange.cs.cornell.edu [128.84.97.8]) by sundial.cs.cornell.edu (8.11.3/8.11.3/M-3.7) with ESMTP id fA83ZtR28236 for ; Wed, 7 Nov 2001 22:35:56 -0500 (EST) content-class: urn:content-classes:message MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" X-MimeOLE: Produced By Microsoft Exchange V6.0.4712.0 Subject: cs615 PAPER 38 Date: Wed, 7 Nov 2001 22:35:55 -0500 Message-ID: <706871B20764CD449DB0E8E3D81C4D4301E7F28D@opus.cs.cornell.edu> X-MS-Has-Attach: X-MS-TNEF-Correlator: Thread-Topic: cs615 PAPER 38 Thread-Index: AcFoBnj3L/Qu8ODkRd2xg37QAGQNYA== From: "Venu Ramasubramanian" To: "Emin Gun Sirer" Content-Transfer-Encoding: 8bit X-MIME-Autoconverted: from quoted-printable to 8bit by sundial.cs.cornell.edu id fA83ZtR28236 Adaptive Protocols for Information Dissemination in Wireless Sensor Networks This paper describes a protocol based on negotiations to disseminate data efficiently throughout the network. Each piece of data is identified with the help of meta-data which is assumed to have a size much smaller compared to the data it describes. Each node advertizes the meta-data of the data it possesses. When neighboring nodes receive this advertisement and realize that they do not posses this data, they send a request to this node and get the data from this node. This 3-way protocol continues throughout the network. Efficiency is achieved because data is transmitted to each node only once. Meta-data transmissons do not increase overhead because they are very small. The above protocol called SPIN is compared with flooding and gossip based broadcast. The gossip protocol considered is quite an inefficient way of using gossip to perform broadcast. A more intellignet way of gossiping such has been done at cornell would produce much better results. The SPIN protocol also only seems to facilitate broadcast of packets. The protocol does not specify how to multicast the data, the difference being that only a few nodes in the network are interested in the data anyway. It is also interesting to note that SPIN does not take advantage of shared communication channel that is typical of ad hoc sensor networks. Thus the same node might send the same packet multiple times to different neighbors. Further the node requesting data must also pick an ideal source to request data in order to get efficiency close to ideal. SPIN 2 is a tweak to SPIN protocol that considers energy available at each node. If a node cannot complete the 3 way protocol it refrains from involving itself with one. It is not very clear how this strategy would increase the lifetime of the network espscially because the energy in the node is already depleted by the time it decides to refrain from participatig. In essence, it is a good idea to use meta-data to improve efficiency of data dissemination in sensor networks. However, i believe that gossip with meta data might prove as efficient as SPIN. From wbell@CS.Cornell.EDU Wed Nov 7 23:02:10 2001 Return-Path: Received: from postoffice.mail.cornell.edu (postoffice.mail.cornell.edu [132.236.56.7]) by sundial.cs.cornell.edu (8.11.3/8.11.3/M-3.7) with ESMTP id fA8428R00648 for ; Wed, 7 Nov 2001 23:02:08 -0500 (EST) Received: from [192.168.1.105] (syr-66-24-16-64.twcny.rr.com [66.24.16.64]) by postoffice.mail.cornell.edu (8.9.3/8.9.3) with ESMTP id XAA24346 for ; Wed, 7 Nov 2001 23:02:06 -0500 (EST) Subject: 615 PAPER #38 From: Walter Bell To: egs@CS.Cornell.EDU Content-Type: text/plain Content-Transfer-Encoding: 7bit X-Mailer: Evolution/0.16.99+cvs.2001.10.18.15.19 (Preview Release) Date: 07 Nov 2001 23:01:47 -0500 Message-Id: <1005192131.1231.0.camel@brute> Mime-Version: 1.0 38) Adaptive Protocols for Information Dissemination in Wireless Sensor Networks This paper presents a family of protocols called SPIN which are used for efficient dissemination of sensor data throughout a energy constrained network. The SPIN protocols are request response style protocols where new information is advertised and interested parties request the data from the source. This reduces overhead by not redundantly sending data when people aren't interested-- all that gets around the network is the advertising packets and the request packets. They add in some additional power conservation abilities such as reducing the amount of request for information when a node becomes power constrained as to not use power for actions it might not complete. They show how these protocols compare to classical flooding and gossiping. Their simulations were not convincing at all-- they implemented a bad gossiping algorithm and asserted it was bad (it redundantly sent to the same person) but didn't fix it, and they neglected to take into account that in flooding, the send overhead is reduced because multiple people can hear the results of one broadcast, effectively killing many birds with one stone. This work begs to be compared to parallel architecture data transmission in high performance message passing machines, which do this type of request meta data transferal to ensure buffering space on a remote node for a message send. It's be interesting to see if some of the research in that area can improve the SPIN ideas, which are geared to be simple, as are transmission mechanisms in parallel machines (as compared to full fledged TCP/IP.) From c.tavoularis@utoronto.ca Thu Nov 8 02:35:32 2001 Return-Path: Received: from bureau6.utcc.utoronto.ca (bureau6.utcc.utoronto.ca [128.100.132.16]) by sundial.cs.cornell.edu (8.11.3/8.11.3/M-3.7) with ESMTP id fA87ZUR21151 for ; Thu, 8 Nov 2001 02:35:30 -0500 (EST) Received: from webmail4.ns.utoronto.ca ([128.100.132.34] EHLO webmail4.ns.utoronto.ca ident: IDENT-NOT-QUERIED [port 57018]) by bureau6.utcc.utoronto.ca with ESMTP id <239671-10688>; Thu, 8 Nov 2001 02:35:23 -0500 Received: by webmail4.ns.utoronto.ca id <164259-210>; Thu, 8 Nov 2001 02:35:03 -0500 To: COM S 615 Subject: 615 PAPER 38 Message-ID: <1005204901.3bea35a518742@webmail.utoronto.ca> Date: Thu, 08 Nov 2001 02:35:01 -0500 (EST) From: c.tavoularis@utoronto.ca MIME-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit User-Agent: IMP/PHP IMAP webmail program 2.2.3 The goal of the proposed routing scheme SPIN (sensor protocols for information via negotiation) is to employ meta-data and resource knowledge to improve the energy and bandwidth efficiency of data collection in sensor networks. It uses node collaboration and data aggregation to improve the accuracy of physical observations and improve robustness against node failure. SPIN addresses problems of classic flooding, which simply stores and forward packets. Implosion is when a node receives multiple copies of the same data from different nodes. Overlap is when overlapping radii forward to the same node. Resource blindness is simply the disregard of resources whose energy is becoming scarce. SPIN uses negotiation between nodes with meta-data to ensure only useful data is forwarded at each node. Meta-data is a type of naming derived from application information and is used for routing decisions. This allows data to be distinguishable at the routing level. Resource adaptation employs a manager that keeps track of a node’s resources and will not commit a node to data aggregation if it is low on energy. Therefore, routing decisions are based on application information, resource state and network topology. Sensor networks are assumed to have sink nodes where users can gather the observations, and SPIN treats all nodes as sinks for fault-tolerance and complete distribution of information. SPIN-1 is a 3-stage handshaking protocol: ADV, REQ and DATA. A node sends out an ADV when it has data to advertise, and nodes that are interested in the ADV return a REQ to request information, such that both messages are defined by informative meta-data and negotiation is applied before every transmission. The final stage is the transfer of the actual DATA. A node acquiring messages of any type can aggregate its own information of the same type before forwarding it on. SPIN-2 builds on SPIN-1 by adding resource-awareness and excluding a node from communication if its energy is below a threshold. SPIN-1 and –2 are contrasted and compared to flooding, gossiping, and ideal forwarding in a small, simulated sensor network. SPIN significantly reduces energy consumption by replacing the forwarding of redundant DATA packets by redundant meta-data packets that are much smaller. SPIN presents a common naming scheme for the transmission protocol and application layers. The unique feature of SPIN is that it de-correlates energy consumption from degree or number of neighbors. This is important because a high degree node lying on a critical path could deteriorate performance. I think SPIN needs to be evaluated and validated in a lossy or mobile environment, especially since the performance of flooding is affected by packet loss and collision. From papadp@ece.cornell.edu Thu Nov 8 10:16:39 2001 Return-Path: Received: from memphis.ece.cornell.edu (memphis.ece.cornell.edu [128.84.81.8]) by sundial.cs.cornell.edu (8.11.3/8.11.3/M-3.7) with ESMTP id fA8FGbR07670 for ; Thu, 8 Nov 2001 10:16:38 -0500 (EST) Received: from kiki.ece.cornell.edu (kiki.ece.cornell.edu [128.84.83.13]) by memphis.ece.cornell.edu (8.11.6/8.11.2) with ESMTP id fA8FFeJ32209; Thu, 8 Nov 2001 10:15:40 -0500 Date: Thu, 8 Nov 2001 10:19:43 -0500 (EST) From: "Panagiotis (Panos) Papadimitratos" To: Emin Gun Sirer cc: Panagiotis Papadimitratos Subject: 615 PAPER 38 Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Review of:"Adaptive Protocols for Information Dissemination in Wireless Sensor Networks," by W. Heinzelman, J. Kulik, H. Balakrishnan The presented protocols aim at supporting efficient data distribution over a (fixed-topology) sensor network so that transmission redundancies are avoided and energy (related to both processing and transmission) is saved. The design of the scheme is data-centric, and data transmissions are regulated at the application layer. Meta-Data, i.e., descriptors extracted from the application-specific data semantics, are used by nodes to advertise the possession of 'fresh' data units and place requests. Upon reception of a (list of) request(s) a node forwards the actual data. In this way, redundant data, i.e., data already seen by a node, not required at any point, or, information that has been acquired by the node itself (e.g., geographical overlapping and same measurements), will not be transmitted. On the other hand, SPIN proposes the use of network-aware applications, and in particular, to take advantage of knowledge of the available resources (energy for example) in order to make efficient/appropriate decisions. However, SPIN does not provide a format for the meta-data and the transmission policy, despite the specific applcition context. Moreover, having every network node advertising every single piece of newly acquired data does not appear as a plausible model. In essence, the scheme substitutes flooding of data packets (in order to achieve simplicity of routing) by flooding of smaller packets, i.e., advertisements. And relies on the application semantics to render itself more efficient. As with the previous schemes, the reasons for MANET routing protocols are not appropriate is not provided (esp. in a setting as the presented experiment). Along the same lines, the comparison with gossiping is not fair, since it could also benefit from application semantics. From gupta@CS.Cornell.EDU Thu Nov 8 10:36:16 2001 Return-Path: Received: from zinger.cs.cornell.edu (zinger.cs.cornell.edu [128.84.96.55]) by sundial.cs.cornell.edu (8.11.3/8.11.3/M-3.7) with ESMTP id fA8FaER09865 for ; Thu, 8 Nov 2001 10:36:14 -0500 (EST) From: Indranil Gupta Received: (from gupta@localhost) by zinger.cs.cornell.edu (8.11.3/8.11.3/C-3.2) id fA8FaE226071 for egs@cs.cornell.edu; Thu, 8 Nov 2001 10:36:14 -0500 (EST) Message-Id: <200111081536.fA8FaE226071@zinger.cs.cornell.edu> Subject: 615 PAPER 38 To: egs@CS.Cornell.EDU Date: Thu, 8 Nov 2001 10:36:14 -0500 (EST) X-Mailer: ELM [version 2.5 PL3] MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Adaptive protocols for information dissemination in wireless sensor networks, Heinzelman, Kulik, Balakrishnan. Reviewer: Indranil Gupta This paper evaluates a simple protocol for information dissemination by broadcasting - using an advertisement+request handshake protocol on links before sending the data over it - against other information dissemination protocols such as flooding, gossiping. The new protocol (called SPIN) appears to consume lesser energy, and deliver more data, flooding and gossiping. If the data stream is only finite (and not continuous), there is not energy dissipation in SPIN after convergence, while the authors' version of gossiping spends energy afterward too. Comments: - The version of gossiping used by the authors is an incomplete design, as admitted by the authors. Adding state to gossiping would not only lower its energy use and increase its data delivery rate, but perhaps make it perform even better than SPIN. - There is another version of (probabilistic) 'gossiping' not evaluated by the authors. Here, a node broadcasts information to all its neighbors, but adjusts the neighborhood set size by changing its transmission power. [Li Li et al] have evaluated the performance of this protocol, and it needs to be evaluated against SPIN. - All the authors' protocols have the following disadvantage. If mobility causes a node that has not received a particular data item, to a locality where it is surrounded by nodes than have already received the item, then this node will never receive the data item. From ranveer@CS.Cornell.EDU Thu Nov 8 11:04:22 2001 Return-Path: Received: from exchange.cs.cornell.edu (exchange.cs.cornell.edu [128.84.97.8]) by sundial.cs.cornell.edu (8.11.3/8.11.3/M-3.7) with ESMTP id fA8G4KR13316 for ; Thu, 8 Nov 2001 11:04:20 -0500 (EST) content-class: urn:content-classes:message MIME-Version: 1.0 Content-Type: text/plain; charset="utf-8" X-MimeOLE: Produced By Microsoft Exchange V6.0.4712.0 Subject: 615 PAPER 38 Date: Thu, 8 Nov 2001 11:04:20 -0500 Message-ID: <706871B20764CD449DB0E8E3D81C4D430232E6AA@opus.cs.cornell.edu> X-MS-Has-Attach: X-MS-TNEF-Correlator: Thread-Topic: 615 PAPER 38 Thread-Index: AcFobwaS0O/vX63dQeKTucowDcAd/Q== From: "Ranveer Chandra" To: "Emin Gun Sirer" Content-Transfer-Encoding: 8bit X-MIME-Autoconverted: from base64 to 8bit by sundial.cs.cornell.edu id fA8G4KR13316 Adaptive Protocols for Information Dissemination in Wireless Sensor Networks This paper proposes energy-efficient, fast converging algorithms, called SPIN, for information dissemination in a sensor network. SPIN attempts to solve 3 particular problems of flooding: those of implosion, overlap, and resource blindnesss. Implosion and overlap are solved by using a 3-way handshake, and algorithms are provided a framework to become resource aware. SPIN protocols are definitely a plus for some cases. If provided with efficient meta-data, and in the presence of big messages, SPIN protocols could possibly perform much better than the alternatives that have been suggested. The idea of resource-awareness is also great for sensor devices that have limited battery power. However, SPIN does raise some issues, such as: 1) No shared communication medium is assumed!!!!! Are the simulations valid then? 2) A 3-way handshake is useful only in the presence of large messages and efficient metadata. Since meta-data are provided by the application, and messages could be of any size, it is doubtful that SPIN would be helpful for all cases. 3) A very bad gossiping algorithm is used to compare SPIN. Additionally, flooding should use broadcast that is the same in energy consumption to a SPIN unicast, but is more effective. It is unclear why the authors chose such inefficient algorithms to compare SPIN. 4) This paper is based on the observation that computation and communication is equally expensive. This seems to be an incorrect assumption, as has been shown in previous years that computation is 3000 times cheaper than communication. 5) SPIN relies on the existence of an underlying beaconing protocol!!! AODV on the other hand could do without any beaconing. It would be helpful to study how much battery power is wasted because of beaconing. From daehyun@csl.cornell.edu Thu Nov 8 11:16:05 2001 Return-Path: Received: from wilkes.csl.cornell.edu (wilkes.csl.cornell.edu [132.236.71.69]) by sundial.cs.cornell.edu (8.11.3/8.11.3/M-3.7) with ESMTP id fA8GG4R14454 for ; Thu, 8 Nov 2001 11:16:04 -0500 (EST) Received: (from daehyun@localhost) by wilkes.csl.cornell.edu (8.9.3/8.9.2) id LAA43816 for egs@cs.cornell.edu; Thu, 8 Nov 2001 11:15:58 -0500 (EST) (envelope-from daehyun) From: Daehyun Kim Message-Id: <200111081615.LAA43816@wilkes.csl.cornell.edu> Subject: 615 PAPER 38 To: egs@CS.Cornell.EDU Date: Thu, 8 Nov 2001 11:15:58 -0500 (EST) X-Mailer: ELM [version 2.4ME+ PL54 (25)] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit This paper presented a family of protocol called Sensor Protocol for Information via negotiation (SPIN), which is designed for energy constrained wireless sensor network. The key ideas of SPIN are negotiation and resource-adaptation. conventional flooding has the following problems; impolsion, overlap and resource blindness. However, SPIN uses negotiation to overcome impolsion and overlap and resource adaptation to overcome resource blindness. SPIN nodes negotiate with each other before transmitting data, so only useful data will be transferred, which eliminates impolsion. SPIN uses naming scheme called meta-data to name the portion of the data that nodes are interested in, which prevents overlap. SPIN nodes poll their resources before data transmission, which allows nodes to cut back on activities when their energy resources are low, which solves resource blindness. Meta-data is a method to descrive data succinctly. It should be smaller than the original data, and disjoint with other meta-data if they represent different original data. The format of meta-data is application specific. So, the interpretation and synthesis meta data are costy. But, the benefit is much bigger. This paper gave simulation results for the five protocols; SPIN1, SPIN2, Flooding, Gossiping and Ideal. SPIN1 uses negotiation and meta-data naming. SPIN2 uses low energy threshold in addition to SPIN1. Flooding sends data to all neighbors. Gossiping selects neighbors randomly and send data to them. Ideal is the optimal case. They measured data transfer rate and energy dissipating rate and showed SPIN is very close to Idea case. Generally, the idea and the methodology are good. But, I think, the comparison targets are too simplified. It will be better if they compare SPIN with other sensor network protocols presented, not just simple schemes. From avneesh@csl.cornell.edu Thu Nov 8 11:40:29 2001 Return-Path: Received: from capricorn.ds.csl.cornell.edu (capricorn.csl.cornell.edu [132.236.71.92]) by sundial.cs.cornell.edu (8.11.3/8.11.3/M-3.7) with ESMTP id fA8GeRR17480 for ; Thu, 8 Nov 2001 11:40:27 -0500 (EST) content-class: urn:content-classes:message Subject: 615 Paper 38 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Date: Thu, 8 Nov 2001 11:42:27 -0500 Message-ID: <97C142C1212ED545B0023A177F5349C40A09C6@capricorn.ds.csl.cornell.edu> X-MimeOLE: Produced By Microsoft Exchange V6.0.4712.0 X-MS-Has-Attach: X-MS-TNEF-Correlator: Thread-Topic: 615 Paper 38 Thread-Index: AcFodFoENx0jwsKaQ2yDsmo9kPJhCg== From: "Avneesh Bhatnagar" To: Content-Transfer-Encoding: 8bit X-MIME-Autoconverted: from quoted-printable to 8bit by sundial.cs.cornell.edu id fA8GeRR17480 Adaptive protocols for Information Dissemination in Wireless Sensor Networks Summary/Critique This paper describes the implementation of protocols for sensor networks (SPIN), which alleviate the implosion, overlap and resource blindness problems of classic flooding and gossip protocols. The idea is to keep the energy usage and communication costs low while optimizing the information disseminated in the network. The SPIN protocols use the methods of negotiation and resource-adaptation to optimize the traffic.Negotiation involves creating meta data information out of the data that the nodes would exchange with each other. This helps in a shorter descriptor for the sensor data, which does not consume a lot of resources. Resource management allows nodes to cut back on certain activities when critical resources such as battery power go low. The SPIN messages are divided into: a. ADV: node advertises with this indicating that it has new information b. REQ: Neighbors of the nodes may request for the above data using this. c. DATA: The actual data. The SPIN suite consists of two protocols SPIN1 and SPIN2; the latter implements the energy-conservation heuristic. The authors then evaluate the above protocols with traditional gossip, and flooding as well as the ideal protocols, and observe favorable results. However there are a few points, that the authors have not made clear: 1. Simulations are done in comparison with a fairly simplistic gossip protocol, this might be a slightly naive implementation. 2. SPIN uses the underlying MAC layer for neighbor discovery, this might also consume energy, I am not user whether that has been taken into account. From viran@csl.cornell.edu Thu Nov 8 11:49:26 2001 Return-Path: Received: from moore.csl.cornell.edu (moore.csl.cornell.edu [132.236.71.83]) by sundial.cs.cornell.edu (8.11.3/8.11.3/M-3.7) with ESMTP id fA8GnOR18563 for ; Thu, 8 Nov 2001 11:49:24 -0500 (EST) Received: from localhost (viran@localhost) by moore.csl.cornell.edu (8.11.3/8.9.2) with ESMTP id fA8GnJk70762 for ; Thu, 8 Nov 2001 11:49:19 -0500 (EST) (envelope-from viran@moore.csl.cornell.edu) X-Authentication-Warning: moore.csl.cornell.edu: viran owned process doing -bs Date: Thu, 8 Nov 2001 11:49:19 -0500 (EST) From: "Virantha N. Ekanayake" To: Subject: 615 Paper 38 Message-ID: <20011108114853.K70528-100000@moore.csl.cornell.edu> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII SPIN is a family of adaptive protocols that purport to route data through a sensor network with lower energy by using data descriptors as a precursor to actual data transmission. Thus, one can avoid flooding the network with large data packets and only send smaller control packets to advertise and request data. Naturally, this does increase point to point data transmission latency since each data transmission requires a three phase handshake. However this is probably not too much of an issue in a sensor network. I'm also not convinced that their energy numbers are accurate -- it seems to me that the fixed energy cost (vs. the energy due to transmitting large packets) of a transmission in a wireless medium (which they did not simulate) would be a more significant factor. It also seems that in a broadcast medium like wireless, the flooding of meta-data packets would create the same amount of contention that data packets would. From teifel@csl.cornell.edu Thu Nov 8 11:56:34 2001 Return-Path: Received: from disney.csl.cornell.edu (disney.csl.cornell.edu [132.236.71.87]) by sundial.cs.cornell.edu (8.11.3/8.11.3/M-3.7) with ESMTP id fA8GuWR19506 for ; Thu, 8 Nov 2001 11:56:32 -0500 (EST) Received: from localhost (teifel@localhost) by disney.csl.cornell.edu (8.11.3/8.9.2) with ESMTP id fA8GuR627537 for ; Thu, 8 Nov 2001 11:56:27 -0500 (EST) (envelope-from teifel@disney.csl.cornell.edu) X-Authentication-Warning: disney.csl.cornell.edu: teifel owned process doing -bs Date: Thu, 8 Nov 2001 11:56:27 -0500 (EST) From: "John R. Teifel" To: Subject: 615 PAPER 38 Message-ID: <20011108113345.B19858-100000@disney.csl.cornell.edu> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII APfIDiWSN: This paper presents SPIN, sensor protocols for information via negotiation, that efficiently communications data across a sensor network. They use a notion of meta-data so that nodes can tag their data. They claim their simulations show that they can extract 60% more data from a sensor network for a given amount of energy, as compared to other approaches. Using the meta-data tags, nodes negotiate about who processes and forwards the data to other nodes. They derive a theoretical ideal network behavior function and show using ns simulations that SPIN converges to this ideal behavior. They present two different versions of the SPIN protocol. They didn't include the loss-prone nature of wireless channels in their simulations, and believe it would not be difficult to do so. I'm not really convinced that this is true and this property may have an impact on their information communication across the sensor network in terms of increased energy usage. From jcb35@cornell.edu Thu Nov 8 14:38:14 2001 Return-Path: Received: from travelers.mail.cornell.edu (travelers.mail.cornell.edu [132.236.56.13]) by sundial.cs.cornell.edu (8.11.3/8.11.3/M-3.7) with ESMTP id fA8JcDR10309 for ; Thu, 8 Nov 2001 14:38:13 -0500 (EST) Received: from travelers.mail.cornell.edu (travelers.mail.cornell.edu [132.236.56.13]) by travelers.mail.cornell.edu (8.9.3/8.9.3) with SMTP id OAA25616; Thu, 8 Nov 2001 14:38:10 -0500 (EST) From: jcb35@cornell.edu Date: Thu, 8 Nov 2001 14:38:10 -0500 (EST) X-Sender: jcb35@travelers.mail.cornell.edu To: egs@CS.Cornell.EDU Subject: 615 PAPER 38 Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII This paper discusses SPIN, an adaptive protocol for information dissemination in wireless sensor networks. To operate efficiently and to conserve energy, the nodes in spin use meta data to describe the data that they collect or need to obtain. SPIN uses a series of advertisement, request for data, and actual data messages to communicate. These messages define a three-way handshake for disseminating data: nodes advertise information, other nodes see this and request the information, and then data can start flowing. They also describe a low energy-conservation version of the SPIN - a node decides if it has enough energy to participate in the proceeding dissemination of information. They evaluate the spin protocol against classic flooding and gossiping, in which a node forwards information to another random node. (I thought this was a rather simplistic way of implementing gossiping). I thought this protocol's well defined operations were a sight for sore eyes after the last paper, even though they didn't seem to take into account reasonable assumptions about the mac layer in their simulations. From mh97@cornell.edu Thu Nov 8 14:47:12 2001 Return-Path: Received: from postoffice.mail.cornell.edu (postoffice.mail.cornell.edu [132.236.56.7]) by sundial.cs.cornell.edu (8.11.3/8.11.3/M-3.7) with ESMTP id fA8JlAR11484 for ; Thu, 8 Nov 2001 14:47:10 -0500 (EST) Received: from mars (dhcp7.csl.cornell.edu [132.236.71.54]) by postoffice.mail.cornell.edu (8.9.3/8.9.3) with ESMTP id OAA21859 for ; Thu, 8 Nov 2001 14:47:09 -0500 (EST) From: "hao ming" To: Subject: 615 PAPER 38 Date: Thu, 8 Nov 2001 14:46:45 -0500 Message-ID: <000301c1688e$18cf5450$3647ec84@mars> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit X-Priority: 3 (Normal) X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook, Build 10.0.2627 Importance: Normal X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2600.0000 adaptive protocols for information dissemination in wireless sensor networks by Wendi Rabiner the point of this paper is that before transfering real data, using meta data to negociate first. the following action is determined by both result of negociation and local resource consumtion situation. this approach can be engery aware, avoiding implosion and overlapping. it makes sense in the case of broadcasting. i am wondering whether it is possible to collaps MAC layer and higher layer so that meta data information can be inserted into the RTS and CTS messages. that would save the bandwidth and energy much. the main drawback of this protocol is that it depends on data length; if the data lengh is comparable to the metat data, it is not good for the SPIN protocol. further, it is not general, only appliable for broadcast mode. -ming