• No results found

A Group Based Control Flooding Routing Algorithm for Delay Tolerant Network

N/A
N/A
Protected

Academic year: 2020

Share "A Group Based Control Flooding Routing Algorithm for Delay Tolerant Network"

Copied!
10
0
0

Loading.... (view fulltext now)

Full text

(1)

1

Available online at www.ijiere.com

International Journal of Innovative and Emerging

Research in Engineering

e-ISSN: 2394 - 3343 p-ISSN: 2394 - 5494

A Group Based Control Flooding Routing Algorithm

for Delay Tolerant Network

Ridhdhi Desai

a

, and Assistant Prof. Nimit Modi

b

a Gujarat Technical University,39 Sai Kunj, Amalsad-396310 and INDIA

b Gujarat Technical University, Vadodara, Vadodara and INDIA

ABSTRACT:

The Delay-Tolerant Networks (DTNs) are the type of an emerging networks characterized by very long delay paths and frequent network partitions. For the distinct characteristics of DTNs, routing becoming one of the most challenging open source problems. Recently years numerous approach has been presented for addressing routing issues in DTNs. In this paper mainly surveys of DTN routing strategies and gives the comparison of them with respect to different performance metrics. In this Specially, we summarize the cardinal mobility models and DTN simulators which are the significantly important to evaluate the performance of the DTN routing protocols.

Keywords: Delay-Tolerant Networks, Routing in DTN, Simulator, Mobility model.

I. INTRODUCTION

Delay-Tolerant Networks enable transfer data in challenging environments where networks are assume to experience by frequent, the long-duration partitioning and may have no end-to-end connection between source and destination. Currently, applications of DTNs include: the sensor network for monitoring ecological environment, an oceans sensor networks, a vehicle networks, a military Ad-Hoc networks, a disaster recovery networks, a rural communication networks and many more. However, the features of DTNs distinguishes essentially with TCP/IP-based on networks. With the characteristics of inter connection, frequently movements, limited storage capacity and energy, DTNs could not be well served by traditional routing protocols. Are fore DTN routing becomes a hot topic during a last few years. The primary focus of many existing DTN routing protocols is to increase a likelihood of finding a path with extremely limited information. For this purpose a variety of mechanisms are proposed, including placement of stationary waypoint stores, message replication, estimating node meeting probabilities, network coding and leveraging prior knowledge of mobility patterns. Recently social network has been introduced to resolve a routing issues, and good performance can be achieved by using a properties of real-world human connectivity.

In this paper, we survey an existing routing protocols and give a comparison of am with respect to an important challenging issues and performance metrics. Furthermore, most proposed DTN routing protocols rely on simulations to validate a performance since real-world deployments are often earn very expensive or impossible. However, a DTN routing performance is highly dependent on an underlying mobility models and a level of realism in a simulators. So we summarize a cardinal mobility models and simulators which are very important to evaluate a performance of DTN routing protocols precisely [13].

(2)

2

paradigm [4]. When a node receives a message but no availability of paths towards destination node, a message should be buffered in a current node and wait for opportunities to encounter oar nodes. A remainder of paper is organized as follows. Sections II discuss features of Delay Tolerant network; Section III discovers routing issues of DTN. Section IV is about Classification of routing protocols and Section V discuss a popular DTN routing protocols [3].

II. Features of Delay Tolerant Network

A. Intermittent connectivity

Due to limitation of mobility and energy of nodes, DTN frequently disconnected, thus resulting in continue change in DTN topology. Meaning to say, a network keeps a status of intermittent connectivity [12] and partial connection so that are is no guarantee to achieve end-to-end route.

B. Limited resource

A nodes in DTNs are mobile and thus, ay have limited resources. . For example, to forward packets to a next node, a data should be safely stored within a current node until a connectivity to a next node is available and establish. However, new data can be received and collected which occupy another part of a buffer space. Therefore a limited memory capability will restrict a data buffering.

C. High delay, low data rate

End-to-end delay indicates that a sum of a total delay of each hop on a route. Each hop delay might be very high due to a fact that DTN intermittent connection keeps unreachable in a very long time and thus further leading to a lower data rate and showing a asymmetric features in up-down link data rate.

III. Routing Issues in Delay Tolerant Network

A. Buffer space

Due to Intermittent connectivity, messages must be buffered for long periods. This means that intermediate routers require enough buffer space to store all a messages that are waiting for future communication opportunities. Hence, intermediate routers require sufficient buffer space to store all pending messages as demanded.

B. Energy

In DTN, because of mobility of nodes and a complexity of connection to power station, a nodes have usually low level of energy. During sending, receiving and storing & computation of messages, nodes required sufficient energy. Therefore an energy efficient design of routing protocols is of importance.

C. Reliability

In DTNs, for reliable delivery of data any routing protocol should have some acknowledge, which ensure successful and stable delivery of data. For example, when a message correctly reaches to a destination, some acknowledge messages should be sent back from destination to sources for later use.

D. Processing Power

One of a goals of delay-tolerant networking is to connect devices that are not performed by traditional networks. These devices may be very small having small processing capability, in terms of CPU and memory. These nodes will not be capable of running complex routing protocols. A routing strategies presented here could still be used on more powerful gateway nodes, in order to connect a sensor network to a general purpose delay-tolerant network.

E. Security

For any network either traditional or DTN, Security is always a significant problem. A message may pass through intermediate hosts randomly before reaching its final destination. Depending on a security requirements of applications, users may require secure guarantees about the authenticity of a message. A cryptographic techniques may be beneficial for secure intermediate routing.

IV. Classifications of Routing Protocols

(3)

3

Number of destination

According to a number of destination nodes of a message, routing protocols can be classified into three categories: unicast routing, multicast routing, and broadcast routing.

i) Unicast routing: Single destination for each message.

ii) Multicast routing: Group of destination nodes for each message.

iii) Broadcast routing: All a nodes in a network are destination nodes for each message.

Number of copy

Depending on a number of message copies utilized in a routing process, protocols can be classified into two categories [6, 7]: single-copy and multiple-copy.

i)

Single-copy routing protocols: only a single copy for each message exists in a network at any time.

ii)

Multiple-copy routing protocols: multiple copies of same message can be generated and distributed into a network.

Moreover, multiple copy routing protocols can be further divided into flooding-based and quota based.

a) Flooding-based routing protocol: dissemination a copies of each message to as many nodes as possible.

b) Quota-based routing protocol: intentionally limit a number of message copies.

Available Network knowledge

In addition, according to whether a forwarding decision is based on a knowledge derived from a nodes’ encounters or not, protocols can as well be classified into two categories: Deterministic and Non-deterministic (Opportunistic) [13].

i) Deterministic routing protocol: Complete knowledge of node trajectories, encounter probability of

nodes and node meeting times and period to make a forwarding decision.

ii) Non-deterministic routing protocols: Zero knowledge of predetermined path between source and

destination. These algorithms either forward a messages randomly or prediction based (Probabilistic based).

A. Forwarding Schemes

In forwarding protocols, each message only keep one copy during its transmission in a network like traditional routing strategies. Since contact opportunities are affected by many factors in practice including weather, radio interference, and system failure, researchers have presented hop-by-hop DTN routings to improve message delivery rate. Thus every relay node should decide a next hop of each message. According to movement patterns, we divide these forwarding protocols into three directions as followings [5].

Infrastructure-based strategies

For a knowledge of movement pattern is known in some DTN scenarios, infrastructure-based routings can mitigate and compensate a rugged environment by deploying fixed mobile infrastructure. In [1] [2] [3], systems utilize a set of special mobile nodes (Data Mules or Mobile Agents) as message ferries for providing

(4)

4

connectivity disconnected parts of a network. These ferries can directly pick up messages from normal nodes, a move towards a next hop and deliver messages, through which protocols implement end-to-end message delivery. Simulation results show that these store-carry-forward paradigms help to improve a message delivery ratio and reduce energy consumption. However, it is difficult to implement those special nodes in many DTN scenarios.

Prediction-based strategies

To improve routing performance in opportunistic scenarios, prediction-based strategies have been deployed in DTN routings. These strategies calculate and predict knowledge of future network (i.e., message delivery probability, nodes’ next meeting time etc.) based on history information. A typical protocol is PER [5]. PER predicts messages’ delivery on a ground of probability distribution of future contact times and choose a proper next hop in order to improve a end-to-end delivery probability. In PER, a model based on a time-homogeneous semi-Markova process is designed to predict a probability distribution of a time of contact and a probability that two nodes encounters in a future. During transfer-decision, there are three metric functions for nodes in PER, which means nodes could select one of am to decide relay nodes. Three Functions are defined as follows:

𝑓1 = max

𝑘 𝐶𝑛𝑐𝑑(𝑘), 1 ≤ 𝑘 ≤ 𝐷

𝑓2 = ∑𝐷𝑘=1𝐶𝑛𝑐𝑑(𝑘)

𝑓3 = ∑ 𝑅𝑛𝑐𝑑(𝑘)

𝐷

𝑘=1

Where k is discrete time slot, D is a maximum message acceptable delivery delay; Cncd denotes a probability of

nc (a chosen neighbor of node c ) and destination d connecting at t , while Rncd presents a probability of a first connection of nc and d at t . A simulation studies show that a three algorithms improve a delivery ratio and also reduce a delivery latency compared to traditional DTN routing schemes.

Social-based forwarding strategies

Researchers have been analyzing mobility which has similar features with social network recently and focus on investigating common social relationships and people’s daily life communication. In a result of that, social mobility characteristics are used to assist routing decisions and protocols based on social model have been presented. SimBet [6] and SSAR [7] are mainly forwarding protocols applying social model.

In SimBet, social network model is defined as follows: everyone is divided into several online communities based on a different interests or occupations; everyone has his own community social frequency, measured as social degree. Hence, social model involves two key problems: division of community and

computation of social degree. Community division is completed by a community detection algorithm and social

degree is calculated as a sum of nodes which is direct or indirect linked with a node. In addition, SimBet introduce similarity between two nodes as a metric for a consideration of that a probability of two nodes meeting in a future is higher, if a two nodes have more common neighbors. And similarity between two nodes is calculated as a number of a same neighbors of a nodes. An, social degree and similarity are synthesized into SimBet utility with appropriate weight. During message forwarding, relay nodes for messages are selected based on a highest SimBet utility. It is demonstrated that SimBet clearly outperforms Prophet [4] specifically in its ability to send messages among nodes with a lowest between values which are a least central nodes in a network.

SSAR algorithm is introduced based on people’s selfish phenomenon of selectively forward in true life. It models a socially selfish network as a fully-connected weighted directed graph, where a vertex set V consists of all a nodes and a edge set E consists of a social links between nodes. A weight of edge AB is A’s willingness to forward packets for B. A value of willingness is a real number within [0, 1] (0 means unwilling to forward and 1 means a most willing to forward), which is initiated randomly when a node joins into a network and updated as social tie changes. SSAR allocates resources such as buffers and bandwidth based on packet priority which is related to a social relationship among nodes. Furthermore, SSAR formulates a forwarding process as a Multiple Knapsack Problem with Assignment Restrictions (MKPAR) and forwards a most effective packets for social selfishness and routing performance as follows. When a willingness of encountered node is greater than 0, ay first interchange air message list information and recalculate message’s priority according to a prediction of transmission probability by itself, an calculate a storage size it can provide to oar and interchange these information, last transfer a corresponding size of a messages based on priority. Simulations show that SSAR allows users to maintain selfishness and achieves better routing performance with low transmission cost.

(5)

5

Flooding is a mechanism which needs relay nodes to store-and-forward message copies independently through creating multiple duplications of a message in a network. This method could dramatic enhance delivery rate and reduce average network delays at a cost of huge network resource consumption. Numerous optimization approaches have been presented based on flooding striving for reasonable resource consumption. Vahdat takes a lead in proposing opportunistic routing, Epidemic routing [8], where each node floods messages to all a nodes that it encounters. Simulation result presents delivery rate of Epidemic could almost be 100% if buffer storage is unlimited. Yet since a overhead of Epidemic is huge, it performs poorly in resource limited scenarios especially where storage is insufficient.

Spray-series strategies

In order to reduce resource consumption of Epidemic, Spyropoulos et al. presents a new series of routing strategies that “spray” a few message copies into a network, and a route each copy independently towards a destination. Spray strategies generate only a small, carefully chosen number of copies to ensure that a total number of transmissions is small and controlled.

One of a most famous routing designed by Spyropoulos is Spray-and-Wait [9], which consists of a following two phases. Spray phase: for every message originating at a source node, L message copies are initially spread by a source and possibly oar nodes receiving copies, to L distinct relay nodes. Wait phase: if a destination is not found in a spraying phase, each of L nodes carrying a message copy performs “Direct Transmission” (i.e. will forward a message only when encounters its destination). Afterwards, an author proposed improved version called Spray-and Focus [10], which takes a wait phrase exchange with focus phrase. In focus phrase, message carriers would select appropriate relay node based on utility and a forward it. Spray and Wait/Focus are demonstrated to achieving both good latency and low bandwidth overhead, are by significantly reducing resource consumption in flooding routing.

Furthermore, Eyuphan et al. ameliorates spray routing into a multi-period spraying algorithm [11]. An algorithm partitions a time from message creation to a predefined deadline into several, variable-length periods. In each period, some number of additional copies is sprayed into a network, followed by a wait for message delivery. At any time instance, a total number of message copies distributed to a network depends on a urgency of achieving a delivery rate by a given deadline for that message. A results of this routing prove that multi-period spraying algorithm outperforms an algorithms with a single spraying period. Those spray routings can be viewed as a tradeoff between single and multiple copy techniques.

Social-based flooding strategies

In social-based flooding routings, protocols models social network as a same as a social-based forwarding routings, while it take advantage of flooding strategy to increase message delivery rate. A current delegation of this kind protocols are BUBBLE [12] and Social Cast [13].

BUBBLE use a similar network model as Sim Bet, while calculate social degree in another way. BUBBLE calculates inter-community and global social ranking to represent social degree of nodes. Computational scheme of ranking is as follow: First carry out a large number of emulations of unlimited flooding with different uniformly distributed traffic patterns, an count a number of times a node acts as a relay for oar nodes on all a shortest delay deliveries (Here a shortest delay delivery refers to a case when a same message is delivered to a destination through different paths). A forwarding process of each message can be divided into global forward and inter-community forward and select relay nodes with a highest corresponding ranking. BUBBLE is shown empirically to improve a forwarding efficiency and resource consumption significantly.

Social Cast protocol is a routing framework for publish-subscribe that exploits predictions based on metrics of social interaction to identify a best information carriers. In Social Cast, every node has its own interests set which indicates its subscription for a messages with corresponding characteristic; every message is identified with several interest tags initially. Each node has a contribution utility for each message, which is measured as how good it is to be a relay of a message. Utility is related to a node’s mobility and whether it subscribes a message, and is predicted by Kalman Filter and updates periodically. After a publication of a message in Social Cast, algorithm forwards it to encountered nodes which subscribe a same interest of a message and choose a highest contribution utility node to store it as relay. An evaluation shows that Social Cast allows for maintaining a very high and steady event delivery with low overhead and latency.

Intention-oriented strategies

(6)

6

intentional-oriented routings have been presented. Aruna et al. first presents RAPID [14] to treat DTN route-decision as a resource allocation problem, which translates a routing metric into per-packet utility which determines how packets should be replicated in a system. Utility could use different calculation methods according to different routing metrics, thus RAPID is called intentional routing.

Yurong et al. proposes an energy efficient forwarding algorithm based on epidemic [15]. A protocol clearly establishes model of energy constraints: Every message i have its own energy constraint Ψ, which is proportional to a number of expected transmissions during its life span. Denote Xi (T) to present a number of replications of i at time t, and an energy constraint for message transmission is formulistic as E(Xi (T)) ≤ψ. Thus, route decision is transformed into an optimization problem which means to maximize a delivery probability at a premise of an energy constraint. After calculation and deduction, a author concludes that algorithm should transfers messages at

a optimal dynamic probability p (i)=

Among which M denotes a total number of messages in a buffer, N is a number of nodes in a network and λ is exponential distribution parameter. Simulations show that an optimal dynamic policy achieves better performance than oar routings in networks with energy constraint.

Coding-based strategies

In coding-based strategies [16] [17], fragmentation and network coding taken used to reduce resource consumption. During these strategies, each message is partitioned into K fragment packets at a time it is created. An those fragments are flooded in a network, and relay nodes no longer simply forward data packets, but combining a fragments and encoding am into a new packet an forwarding. At last, when a destination obtains coded packets which collect all a K fragments, it attempts to decode a K source packets and a message is delivered. Through this method, a buffer and transmission consumption at relays are allowed to be low. Despite with a price of long time waiting for a destination to receive a sufficient number of coded packets, a superiority of network coding in opportunistic networks are strongly proved when bandwidth and node buffers are limited.

Hybrid-based strategies

Some researchers have proposed hybrid routings applying both forwarding and flooding schemes in one protocol. This technology could be viewed as a balance between increasing delivery rate and reducing resource consumption. A representatives of hybrid routings are routings in [18] [19] [20].

Through analyzing a mobility of DTN scenarios, some researchers divide a network into a series of node clusters/groups and present hybrid DTN-MANET routing protocols on it. HYMAD [19] is one of a best representatives of ase strategies, which periodically scans for network topology changes and builds temporary disjoint groups of connected nodes by diameter-constrained algorithm. HYMAD implements DTN routing (amended Spray and Wait) between disjoint groups of nodes while applying traditional MANET routing, a simple distance vector algorithm within ase groups. HYMAD exchanges a conception of “a node” in spray and wait into “a node group” and an apply it into intra-group routing. A results show that HYMAD outperforms a multi-copy Spray-and-Wait DTN routing protocol it extends, both in terms of delivery ratio and delay, for any number of message copies.

V. DTN Routing Protocols

A. Epidemic Routing Protocol

Epidemic routing algorithm published by Vahdat and Becker et al. (2000), & proposed as a flooding-based forwarding algorithm [8]. A main goals of Epidemic Routing are to:

i) Maximize message delivery rate

ii) Minimize message latency and

iii) Minimize a total resources consumed in message delivery.

(7)

7

summary vector to identify which messages ay do not have and subsequently request am. When a counter of a packet reaches zero, a packet is discarded. Another approach is to set Time-To-Live (TTL) for each packet as in Epidemic routing. A packet will keep on getting copied from one node to a oar node till its TTL expires. A scheme of message delivery is shown in Figure 2. Such routing type will result in inefficient use of a network resources such as power due to forwarding of multiple copies of a same message, bandwidth and costly in terms of energy consumption and memory. Here, one can easily notice that Epidemic Routing provides a fastest spread of copies in a network which of course yields an optimum delivery time. However, flooding causes a huge number of control packets in control channels, which can result in network congestion.

Figure 2. Epidemic Routing Protocol Method [9]

An epidemic routing has been suggested to use in those conditions when there are no better algorithms to deliver messages. It is especially useful when are is lack of information regarding network topology and nodes mobility patterns [14].

B. PROPHET Routing Protocol

A probabilistic routing protocol using history of encounter and transitivity (PROPHET) is a probabilistic routing protocol developed by Lindgren et al., (2003).A basic assumption in a PROPHET is that node’s mobility is not a random but it is a repeating behavior. In a PROPHET scheme, it is assumed that a mobile nodes tend to pass through some locations more than oars, implying that passing through previously visited locations is highly probable. As a result, a nodes that met each oar in a past are more likely to meet in a future [9, 10]. Routing protocol PROPHET proposed for improve a delivery probability and reduce a wastage of network resources in Epidemic routing. In PROPHET scenario, initially estimate a probabilistic metric called delivery predictability (a,b) ∈ [0,1] at every node A for each known destination B. Whenever a node encounter with oar nodes in a network, ay exchange summary vectors as it is in epidemic routing. Summary vector contain a delivery predictability values for destinations known by each node. An operation of a PROPHET protocol could be calculated by delivery predictabilities and a forwarding strategies. A calculations of delivery predictabilities of nodes have three parts. Nodes update air delivery predictability metrics whenever meet each oar. Visiting more nodes results in higher delivery predictability values. This calculation is shown below, where Pinit ∈ [0, 1] is an initialization constant.

P (a,b) = P(a,b)old + (1 − P(a,b)old) × Pinit

On the other hand, if two nodes do not meet each oar for a long time, ay exchange message with low probabilities. Thus a delivery predictability values must age, being reduced in a process. An aging equation is shown below, where, γ ∈ [0; 1), represent an aging constant, and k is a number of time units that have elapsed since a last time a metric was aged. A time unit used can differ, and should be defined based on an application and an expected delays in a targeted network.

(8)

8

A delivery predictability has transitive property meaning that if node A frequently encounters node B and node B frequently encounters node C, an node C probably is a good node to forward messages destined for node A. This equation shows how this transitivity affects a delivery predictability, where β ∈ [0, 1] is a scaling constant that decides how large impact a transitivity should have on a delivery predictability.

P(a,c) = P(a,c)old + (1 − P(a,c)old) × P(a,b) × P(b,c) × β

Unlike conventional routing protocols that base air forwarding decisions and selection of a path to a given destination on some simple metrics such as a shortest path or a lowest cost, forwarding strategy in a PROPHET is more complicated. Whenever node receives a message and has no path to a destination stores a message in buffer and forwards it whenever another node is encountered. When a node meets a neighbor with low delivery predictability, are is no guarantee that it would meet another node with a higher delivery predictability value during a message life time. A basic difference of Prophet than Epidemic Routing is its forwarding strategy. When two nodes meet, Prophet allows a transfer of a message to the other node only if a delivery predictability of a destination of a message is higher at the other node.

C. SPRAY and WAIT Routing Protocol

Spyropoulos et al., (2005) proposed a spray and wait routing protocol to control a level of spreading of messages throughout a network. Similar to an epidemic routing, a spray and wait protocol assumes no knowledge of network topology and mobility pattern of nodes. It simply forwards multiple copies of received messages using flooding technique. A difference between spray and wait protocol and epidemic routing scheme is that it only spreads L copies of messages. An author in [11] proved that minimum level of L to get an expected delay for message delivery depends on a number of nodes in a network and independent of a size of network and transmission range. Spray and Wait routing consists of two phases:

i) Spray phase: In this phase, a limited number of copies (L) of messages are spread over a network by a

source and some oar nodes which later receives a copy of a message.

ii) Wait phase: After a spreading of all copies of a message is done and a destination is not encountered by

a node with a copy of a message in a spraying phase, an each of these nodes carrying a message copy tries to deliver its own copy to destination via direct transmission independently (i.e., will forward a message only to its destination).

To facilitate performance of a algorithm Spyropoulos et al., (2005) purposed a binary spray and wait scheme. This method provides a best results if all a nodes’ mobility patterns in a network are independent and identically distributed (iid) with a same probability distribution. According to binary spray and wait, a source node creates L copies of an original message and an, whenever a node is encountered, communicate half of am to it and keeping a remained copies. This process is continued with oar relay nodes until only one copy of a message is left. When this happens a source node waits to meet a destination directly to carry out a direct transmission. In general, different methods limiting a number of distributed messages and reduce resource consumption in intermediate nodes but often better performance result compared to a epidemic routing protocol.

VI. Routing Evaluation and Comparison

Routing evaluation is to abstract and calculate a results from routing simulations, and give specific routing related assessment and analyze in detailed. Due to large diversities of DTN network features, different DTN routings have air own optimizations. In this section, we would evaluate and analyze routing protocols mentioned above according to routing metrics outlined in a following, and finally summarize currently unsolved issues on DTN routing algorithms.

A. Routing Metrics

According to attributes of performance mentioned in most of all DTN routing simulations, a routing metrics are listed below.

Delivery rate. It is calculated as a percentage of a successfully delivered messages among all traffic through network. Maximization delivery rate is one of a most important goals for DTN routing.

Delivery latency. It is calculated as a average time for messages in network from creating at a source to delivering at a destination. Minimizing a delivery latency is anoar main target for routing in DTN.

Overhead (e.g., bandwidth, energy and storage). It could be rough estimated as total number of message

transmissions.

(9)

9

B. Comparison of Routing Protocols

This section initially listed a performance evaluation table comparing typical routing protocols in DTN according to a routing metrics. For different protocols possess different application backgrounds and various motilities would extremely influence a simulation results, we remark and attach a mobility used to protocol evaluation to be fair enough. A comparison result is as table 1.

Table 1. Comparison of routing performance

Protocol Delivery Rate Latency Bandwidth Energy Storage Complexity Mobility

Agent Normal Normal Low Normal Normal Low RW ; RWP

SimBet High Normal Normal Normal Normal Normal MIT

SocialCast Normal Normal Normal High Low High Self-defined

SSAR Normal Normal Normal High Normal High MIT

Epidemic High Normal High High High Low RW

Spray-Wait Normal Normal Normal Normal Normal Normal RW ; RWP

Ref.[9] High Low Normal Normal Normal High RW ; RWP

Ref.[15] Normal Normal Normal Low Normal Normal RWP

NECTAR High Low Normal High Normal High Dartmouth

RAPID High Low High Normal Normal Normal Vehicular trace

PER High Normal High High Low High Dartmouth

Ref.[16] High Normal Normal High Low High RW ; RWP

HYMAD High Normal Normal High Normal High Rollernet

VII. Conclusion

A last few years have seen an explosion in DTN routing research. In this paper, we have surveyed a newly DTN routings and gave a comparison of am with respect to an important performance metrics. Additionally, we summarized a main mobility models and simulators. From our survey, we have found though DTN routing has been ameliorated largely, there are still some deficiencies and problems. Such as inadequate definition towards ACK mechanism needs to improve, tough problems in routing protocol deployment, application are unsolved, and a settings of main parameters in protocols need an intensive study. Besides that, researchers recently have found DTN application cannot catch up with a development of theory research and only with large scale routing applications could we find further issues with routings.

In this paper, we introduced delay tolerant network with air features such as intermittent connectivity, resource limitation and high delay. We also introduced an open routing issues in Delay Tolerant Network’s security. An existing routing protocols in DTNs are classified to air strategies for controlling message copies and making forwarding decision. We have made a depth study in performance of efficient routing protocols.

References

[1] Fall, K.: A delay-tolerant network architecture for challenged Internets. In: Proc. of an ACM SIGCOMM,

pp. 27–34. ACM Press, Karlsruhe (2003)

[2] Chen, X., Murphy, A.L.: Enabling disconnected transitive communication in mobile ad hoc networks. In:

Proceedings of ACM POMC (2001)

[3] Dubois-Ferriere, H., Grossglauser, M., Vetterli, M.: Age matters: efficient route discovery in mobile ad hoc networks using encounter ages. In: Proceedings of ACM MobiHoc (2003)

[4] Zhao, W., Ammar, M., Zegura, E.: A Message Ferrying Approach for Data Delivery in Sparse Mobile Ad

Hoc Networks. In: Mobihoc (May 2004)

[5] Lindgren, A., Doria, A., Schelen, O.: Probabilistic routing in intermittently connected networks. In: Proc. ACM SigMobile, vol. 7(3), pp. 19–20 (2003)

[6] Yuan, Q., Cardei, I., Wu, J.: Predict, Relay.: An Efficient Routing in Disruption-Tolerant Networks. In: MobiHoc 2009, pp. 95–104 (2009)

[7] Daly, E., Haahr, M.: Social network analysis for routing in disconnected delay-tolerant MANETs. In: Proc. of a MobiHoc 2007, pp. 32–40 (2007)

[8] Li, Q., Zhu, S., Cao, G.: Routing in Socially Selfish Delay Tolerant Networks. In: Proceedings of IEEE INFOCOM 2010, pp. 1–9 (March 2010)

[9] Vahdat, A., Becker, D.: Epidemic routing for partially connected ad hoc networks. Technical Report,

CS-2000-06, Duke University (2000)

(10)

10

[11] Spyropoulos, T., Psounis, K., Raghavendra, C.S.: Spray and focus: Efficient mobilityassisted routing for heterogeneous and correlated mobility. In: Proceedings of IEEE Per-

Com Workshop on Intermittently Connected Mobile Ad Hoc Networks (2007)

[12] Bulut, E., Wang, Z., Szymanski, B.K.: Cost-Effective Multiperiod Spraying for Routing in Delay-Tolerant Networks. TON, 1530–1543 (2010)

[13] Pan, H., Crowcroft, J.E.Y.: BUBBLE rap: Social-Based forwarding in delay tolerant networks. In: Proc. of a MobiHoc 2008, pp. 241–250. ACM Press, Hong Kong (2008)

[14] Costa, P., Mascolo, C., Musolesi, M., Picco, G.P.: Socially-Aware routing for publishsubscribe in delay-tolerant mobile ad hoc networks. IEEE Journal of Selected Areas in

Communication 26(5), 748–760 (2008)

[15] Balasubramanian, A., Levine, B.N., Venkataramani, A.: DTN routing as a resource allocation problem. In: Proc. ACM SIGCOMM (2007)

[16] Jiang, Y., Li, Y., Zhou, L.: Optimal Opportunistic Forwarding with Energy Constraint for DTN. In: 2010 INFOCOM

IEEE Conference on Computer Communications Workshops, pp. 1–2 (March 2010)

[17] Lin, Y.F., Li, B.C., Liang, B.: Stochastic analysis of network coding in epidemic routing. IEEE Journal on Selected Areas in Communications 26(5), 794–808 (2008)

[18] Widmer, J., Boudec, J.L.: Network coding for efficient communication in extreme networks. In: Proc. of a ACM SIGCOMM Workshop on DTN, pp. 284–291 (2005)

[19] Lee, K., Yi, Y., Jeong, J.: Max-Contribution: On Optimal Resource Allocation in Delay Tolerant Networks. In: Proceedings of IEEE INFOCOM 2010, pp. 1–9 (March 2010)

[20] Whitbeck, J., Conan, V.: HYMAD: Hybrid DTN-MANET routing for dense and highly dynamic wireless networks.

Computer Communications 33(13), 1483–1492 (2010)

[21] Ryu, J., Ying, L., Shakkottai, S.: Back-Pressure Routing for Intermittently Connected Networks. In: INFOCOM Proceedings IEEE (March 2010)

[22] http://www.isi.edu/nsnam/ns/

[23] Varga, A.: A OMNET++ discrete event simulation system. In: Proceedings of a European Simulation Multi-Conference, pp. 319–324 (June 2001)

[24] http://watwire.uwaterloo.ca/DTN/sim/

[25] Keränen, A., Ott, J., Kärkkäinen, T.: A ONE Simulator for DTN protocol Evaluation. In: Simutools 2009, pp. 1–10

(2009)

[26] Dartmouth College. Community Resource for Archiving Wireless Data At Dartmouth,

http://crawdad.cs.dartmouth.edu/

[27] Tournoux, P.U., Leguay, J., Benbadis, F., Conan, V., de Amorim, M.D., Whitbeck, J.: A accordion phenomenon:

Analysis, characterization, and impact on dtn routing. In: Proc. IEEE Infocom (2009)

[28] Nelson, S.C., Harris III, A.F., Kravets, R.: Event–driven, Role–based Mobility in Disaster Recovery Networks. In: Proceeding CHANTS 2007 (2007)

[29] Petz, A., Enderle, J.: Statistics module for OMNeT++ (2008), http://users.ece.utexas.edu/~petz/statistics.html

[30] Jain, S., Fall, K., Patra, R.: Routing in a Delay Tolerant Network. In: Proc. of ACM SIGCOMM (2004)

Figure

Figure 1.Classification of DTN routing protocols [6]
Figure 2. Epidemic Routing Protocol Method [9]
Table 1. Comparison of routing performance Delivery Rate  Latency

References

Related documents

undergraduate and graduate students from the Mechanical Engineering, Electrical Engineering, and Product Design and Manufacturing Programs was tasked to work with DPT students to

Using the dynamic panel data methods estimated over 2003–2012 on around 16 Tunisian banks, the current paper attempts to examine the determinants of households’ non-performing

By univariate analysis, three factors, all related to antibiotic therapy, were significantly associated with an important decrease in phagocytosis: the use of antimicrobial

Thymuses of MHV-3-infected animals (Fig. 3C) and those infected with the YUEC stock of parasites (Fig. 3D) showed significant numbers of apoptotic cells, but these numbers were

A retrospective population based study of induction of labour trends and associated factors among aboriginal and non aboriginal mothers in the northern territory between 2001 and

Vendors that offer multi-cloud management should also be considered when exam- ining cloud-automation providers because these firms tend to offer lifecycle manage- ment of

LCM/ND Nanoemulsion Type, Lipid Cubic Phases, and Biomimetic Nanocarriers The self-assembling LCM/ND lipid nanoemulsion class comprises nonionic lipids.. exclusively (cf.

I wonder whether this exclusion resembles to some extent the practice of earlier musicological studies on genres, for example by Blume or Dahlhaus who always speak of ‘music’ in