Internet Connectivity for Mobile Ad Hoc Networks Using Hybrid Adaptive Mobile Agent Protocol

Transcription

1 The International Arab Journal of Information Technoloy, Vol. 5, No. 1, January Internet Connectivity for Mobile Ad Hoc Networks Usin Hybrid Adaptive Mobile Aent Protocol Velmuruan Ayyadurai 1 and Raaram Ramasamy 2 1 Department of Electronics Enineerin, Anna University, India. 2 Department of Information Technoloy, Thiaaraar Collee of Enineerin, India. Abstract: Internet-based Mobile Ad Hoc Networkin (MANET) is an emerin technoloy that supports self-oranizin mobile networkin infrastructures. This is expected to be of reat use in commercial applications for the next eneration internet users. A number of technical challenes are faced today due to the heteroeneous, dynamic nature of this hybrid MANET. A new hybrid routin scheme AODV ALMA is proposed, which act simultaneously combinin mobile aents to find path to the ateway and on-demand distance vector approach to find path in local MANET is one of the unique solution. An adaptive ateway discovery mechanism based on mobile aents makin use of pheromone value, pheromone decay time and balance index is used to estimate the path and next hop to the ateway. The mobile nodes automatically confiure the address usin mobile aents first selectin the ateway and then usin the ateway prefix address. The mobile aents are also used to track chanes in topoloy enablin hih network connectivity with reduced delay in packet transmission to Internet. The performance tradeoffs and limitations with existin solutions for various mobility conditions are evaluated usin simulation. Keywords: Hybrid MANET, mobile aents, AODV ALMA, adaptive ateway discovery. Received April 7, 2006; Accepted December 1, Introduction A Mobile Ad Hoc Network (MANET) is an autonomous network that can be formed without any established infrastructure. It consists of mobile nodes equipped with a wireless interface that are free to move and establish communication on the fly. The mobile nodes act as hosts and routers, havin routin capabilities for multi-hop paths connectin nodes which cannot directly communicate. The idea of facilitatin the interation of MANET and fixed IP network i.e., Internet has ained lot of momentum within the research community recently [11]. This kind of interation forms the basis for the next eneration internet usae in which mobile nodes are expected to provide service information any time, any where for ubiquitous access. These kinds of networks are named as hybrid MANET which can be deployed quickly and easily. The potential applications are like conference meetins, wireless classrooms, vehicular operations in a hihway, ubiquitous information processin in train, and airport. Researchers presently make use of the social behaviours of insects and of other animals for developin a variety of different oranized behaviours at the system level. In particular, ants have inspired a number of methods and techniques amon which the most studied and the most successful is the eneral purpose optimization technique know as Ant Colony Optimization (ACO). ACO takes inspiration from the forain behaviour of some ant species. These ants deposit pheromone on the round in order to mark some favourable path that should be followed by other members of the colony. Because of the robustness and efficiency of ACO they have recently become a source of inspiration for the desin of routin alorithms. Followin the ACO principles, Ant like Mobile aents (ALMA) are simple packets that explore the network and collect data for routin [2, 3]. They communicate with one another indirectly by exchanin routin information by writin and readin data to from routin tables. This is similar to ants exchanin food information with each other. By usin mobile aent exploration, the idea is to put the intellience across the network and make the routin distributed with hih connectivity. The important constraint in usin the mobile aents for routin is that the overhead is proportional to the number of aents present in the network. This in return reduces the overhead compared with traditional routin protocols that use frequent updates and reactive updates. Fiure 1 shows the architecture in which the mobile nodes form local MANET access with multi-hop wireless links to provide service within them. The Internet Gateway (IGW) nodes form lobal access to provide internet service to mobile nodes. The main technical challenes faced today in this scenario are Internet ateway discovery, address autoconfiuration and reachin the destination. The ACO mechanism is used as an adaptive reactive ateway discovery approach interated with address auto-

2 26 The International Arab Journal of Information Technoloy, Vol. 5, No. 1, January 2008 confiuration. The new hybrid scheme, proposed in this paper, combines local access with lobal access, named as AODV ALMA. The remainder of the paper is oranized as follows. The paper ives an overview of the related work in section 2. The new protocol description is iven in section 3. The performance analysis is simulated and compared in section 4. The future enhancement is iven as conclusion in section 5. IGW1 Internet IGW2 Fiure 1. Mobile ad hoc network connectivity with Internet. 2. Related Existin Proposals IGW3 Mobile Node Wireless Link The first proposal by Broch et al. [2] is based on interation of Mobile IP and MANET, employ a source routin protocol. The second proposal MIPMANET [6] followed a similar approach based on AODV, but works only with Mobile IPv 4 approach, because it requires Forein Aents (FAs). The proposal by Ammari et al. [1] ives a mobile ateway approach based on DSDV protocol and analyzed the performance. In Ratanchandani et al. [8] proposal, a hybrid scheme for propaatin the advertisements up to a certain number of hops is iven. The mobile nodes located out of this propaation scope will reactively find the ateways when needed. The performance of this approach depends on the Time-To-Live (TTL) value, set for particular scenario and network conditions under consideration. In Lee [5] proposal, another more sophisticated hybrid approach the advertisements are sent out only when chanes in topoloy are detected. However, they rely on a source based routin protocol, which limits the applicability to particular type of routin protocol. In Ruiz et al. [9] proposal, an adaptive ateway discovery mechanism that outperforms existin hybrid approaches is proposed. The key scheme is that TTL value for proactive ateway advertisements is adusted dynamically to network conditions. The mobile nodes out of the propaation scope reactively find the ateways. The proposals available till date focused only on ateway discovery and are not a complete solution for inter-workin with fixed IP networks. The complete proposals available to interate both auto confiuration and ateway discovery are namely Wakikawa [12], Jeler [4] and Sinh et al. [11]. In Wakikawa [12], mobile nodes discover the ateway either reactively or proactively to confiure their IP address automatically. In the proactive approach, ateway periodically floods the ateway advertisement messaes up to certain number of hops in the MANET. In the reactive approach, mobile nodes flood the ateway solicitation messae to all ateways reachable. The ateway receivin the ateway solicitation will unicast back the ateway advertisement to the mobile nodes. The ateway advertisement messae contains the lobal IPv 6 address of the ateway, the network prefix advertised by the ateway, the prefix lenth and the lifetime associated with this information. In Jeler et al. [4] proposal, proactive approach is used in which the ateways advertise themselves by periodically floodin ateway information messaes. The floodin of these messaes is based on the idea of prefix continuity. A mobile node selects the best path towards the ateway usin one of the metrics such as distance, stability, or delay from all the ateway information messaes received. By selectin the ateway information messaes, the mobile node confiures an IPv 6 address based on the selected prefix. This approach does not require the use of routin headers In Sinh et al. [11] proposal, ateways act as nodes that are one hop away from the access router. The access routers are equipped with a wireless interface. The first node becomin a ateway is called the default ateway, and it is responsible for sendin out periodic ateway advertisements. The other ateways for a iven router are called candidate ateways. A candidate ateway becomes a default ateway when it stops receivin ateway advertisements from the default ateway for some time. It defines ateway selection based on bandwidth balancin, but unfortunately bandwidth parameters and their use not described in the current version of the specification. 3. Proposed Protocol Description: AODV ALMA In hybrid MANET network, connectivity of mobile nodes with ateway nodes and mobile nodes with other active mobile nodes is uncertain due to the dynamic network topoloy. The mobility of mobile nodes cause more delay in findin the route to destination either in Internet or in local MANET. The newly proposed hybrid mechanism AODV ALMA overcomes this by combinin the on demand distance vector approach with ACO. The on demand distance vector approach, i.e., AODV used to find the path in local MANET. The ACO approach, i.e., ALMA (Ant like Mobile Aents) used to find the path

3 Internet Connectivity for Mobile Ad Hoc Networks Usin Hybrid Adaptive Mobile Aent Protocol 27 to the ateway. The hybrid mechanism operates simultaneously when a route to the destination is needed by the source mobile node. The routin table structure of mobile node is dynamically updated on demand usin the proposed hybrid mechanism is shown in Table 1. The ateway contains two routin tables, one to update information in local MANET and another to update the information from Internet is shown in Table 2. Table 1. Routin table in mobile nodes. Destination address Next Hop address Number of Hops Sequence number for the destination Active neihbours for this route Expiration time for the route table entry Pheromone value Pheromone decay time Balance index Table 2. Routin table in ateway for local MANET. Source address Next Hop address Number of Hops Sequence number for the source Active neihbours for this route Expiration time for the route table entry Pheromone value Pheromone decay time Balance index After ettin the route to the destination in the mobile node usin the proposed hybrid mechanism, data packets can be forwarded. The data packets forwarded from the mobile nodes can be either to the Internet or to the local MANET usin the routin table entries available reactively. The data packets to the destination in Internet are forwarded to the ateway usin the entries created by ACO mechanism such as number of hops, next hop, pheromone values, and pheromone decay time with balance index. The data packets to the destination in the local MANET are forwarded usin the entries created by AODV such as number of hops, next hop, and active neihbours for this route with expiration time. The data packets follow the IPv 6 format for transmission. The main reason for usin ACO in ateway discovery is to set the link adaptive parameters based on the current topoloy of the links traversed by the mobile aent packets in particular time. This is similar to the way ants walkin to and from a food source deposit on the round a substance called pheromone. The other reason for usin ACO is to take the decision in selectin the stable path to the ateway based on the pheromone concentration on the current node provided for each possible link. This will provide service access to the mobile node and next hop to the ateway with hih network connectivity for a lon period of time. The ACO principle in a unifyin framework with the on demand distance vector approach is the first approach of this kind for hybrid MANET access Gateway Discovery in AODV ALMA The proposed protocol is used to discover paths and to maintain paths in local MANET for dynamic topoloy chanes. The mobile nodes initially use the available temporary address for broadcastin before confiurin the address based on the ateway. The mobile nodes flood ACO messaes and AODV messaes simultaneously to discover routes either in lobal Internet or in local MANET. The ACO messae first sent to discover the ateway is the Forward Routin Mobile Aent (FRMA) messae. The FRMA messae is multicast with the all manet ateway multicast address. The intermediate mobile nodes in the path to the ateway receive the FRMA messae and mark the messae with its node information into the history field. The FRMA messae on traversin the link throuh the intermediate nodes to the ateway will et the link stability factor in terms of the pheromone value and pheromone decay time. The intermediate node after markin the FRMA messae will rebroadcast in local MANET until it discovers the ateway. The FRMA messae keeps the list of mobile nodes [1, 2,, n] it has visited in the history field. The Route Request Messae (RREQ) of AODV is used to find the path in local MANET. The FRMA packet updates the routin table entries of active intermediate mobile nodes with link probability value in terms of the pheromone value with decay time. The ateway detectin the FRMA packet transfer the history field with the total link probability value and ets converted as Backward Routin Mobile Aent (BRMA). The pheromone value in the FRMA packet is bound to decay with time and number of hops travelled. The decision process is that larer the value of pheromone, then shorter the time taken to reach the ateway, but smaller the value of pheromone loner the time to reach the ateway. The BRMA packet has ot hiher priority over FRMA packet in traversin the links in the reverse direction. The BRMA packet is unicast from the ateway to the source mobile node will pass throuh the list of intermediate nodes visited by the FRMA packet. The BRMA packet updates and verifies the routin table entries created by the FRMA packet by findin the topoloy chanes affectin the link. The BRMA packet contains the followin information, the lobal IPv 6 address of the ateway, the network prefix of the ateway, the prefix lenth, lifetime, pheromone value, pheromone decay time with the balance index. After reachin the source mobile node the BRMA packet

4 28 The International Arab Journal of Information Technoloy, Vol. 5, No. 1, January 2008 update the routin table with the information collected for the link traversed. The source mobile node selects the stable path to the ateway based on the total link probability towards the ateway usin the pheromone value and the pheromone decay time detected by the mobile aents. The balance index will ive the load in the ateway based on the maximum number of mobile nodes active to the number of mobile aents present in the network. The balance index value lesser means minimum the load in the ateway. To forward the data packets mobile node has to first select the ateway based on the stable path and next with the balance index information updated by BRMA packets. The proposed protocol is adaptive since it reinforces the route to the ateway information based on the link probability value with time usin mobile aents. The source mobile node receivin more than one BRMA packet selects the adaptive ateway by ettin maximum link probability value with lesser balance index. The source mobile node auto-confiures its IP address after selectin the ateway usin the prefix of the ateway with the suffix wireless interface address. The active intermediate mobile nodes also maintain the adaptive information set by the BRMA packet before the link probability reaches minimum threshold value over time. On reachin the threshold value the active intermediate mobile node enerate a new FRMA packet with the source mobile node address available from its routin table. The enerated FRMA packet updates the route to the ateway with either the new active node or the old active node for forwardin the data packets. The FRMA and BRMA packets explore and quickly reinforce the paths to the ateway adaptively. They also ensure the previously discovered paths do not et saturated when the actual link fails. The flowchart in Fiure 2 shows the process the mobile node uses the mobile aents to discover the route to the ateway. The proposed hybrid scheme will reduce the route discovery latency with less converence time Analytical Model The analytical model ives the mappin of ACO to find routes to the ateway in hybrid MANET. It assumes a Fixed Radius Model as the topoloy model with R the mobile nodes communication rane. The mobile node find route to the nearest ateway for internet access to prevent mobile node isolation usin multi-hop communication to the ateway. The density and radius must be lare enouh to keep the proportion of mobile nodes which do not belon to the eoraphical area very less. The proposed protocol use mobile aents to find the path between the source mobile node vs and the ateway node v is mapped to the connected raph ( V, E) G= with n= V nodes. The path lenth in G is iven by the number of nodes on the path. Each ede in G, i.e., e( ) E, connectin the nodes v has a pheromone value vi and ϕ which is modified usin mobile aents visitin the intermediate nodes. The pheromone concentration ϕ is an indication of the usae of this ede. Node has data to send (Internet Host) FRMA packet created to explore routes to ateway usin ALL- MANET-MULTICAST address Broadcast Copies of FRMA to neihbors usin normal queues Initialize INFO_MAP stack, Pheromone value, Pheromone Decay No Active node Forward FRMA packet No Sourc e node Yes Drop FRMA packet Yes Update FRMA INFO_MAP stack, Pheromone value, Pheromone decay time Waitin for BRMA packet Fiure 2. Process in mobile source node usin AODV ALMA. In order to forward a data packet towards the ateway node v, the forwardin equation is formulated in the proposed protocol to ive link probability usin pheromone value present between neihbourin nodes vi and v. The equation (1) maps the total probability of the link towards the ateway v with pheromone on each outoin link v i at node v as p,. The mobile aent located in node v i use pheromone value of each link to the ateway as ϕ,. The specific next hop neihbour is chosen accordin to the probability distribution in each link.

5 Internet Connectivity for Mobile Ad Hoc Networks Usin Hybrid Adaptive Mobile Aent Protocol 29 p, N = = 1 0 F ( ϕ, + K) if N i i F ( ϕ + K) (1), if N The forwardin Equation 1 ives Ni the number of neihbours to node i, the constant F and K is used in tunin the routin behaviour of the proposed protocol. The pheromone threshold value K, determines the sensitivity of the link probability calculations to small amounts of pheromone. If K 0 larer, then lare amounts of pheromone will be present before an appreciable effect will be seen in the link probability. The pheromone sensitivity value F 0, used to modulate the differences between pheromone amounts present in link probability. If the value of F>1 will accentuate differences between links, while F<1 will deemphasize them and F=1 ives a simple normalization form. The link probability p, of the node v i fulfils the constraint in ivin transition probabilities as in Equation 2. p, = 1, i [ 1, N] Ni i (2) Durin the route findin process to the ateway mobile aents deposit pheromone on the edes. The pheromone concentration is varyin from one ede to another between the nodes connectin them by an amount ϕ. The pheromone value chane at the ede ( ) e v v when mobile aents movin from node vi to node v are iven in Equation 3. ϕ = ϕ + ϕ (3),, The pheromone value ϕ, can be used to compute the link probability p, of the node v as the next hop to reach the ateway node from node v i. A link connectin nodes vi and v is added to the raph as l and this path is selected to the ateway in the proposed protocol. The decision ives the fact that the Euclidean distance between the nodes connectin the ateway is less thanr and the path selected is stable with the threshold value set. The condition assumes the case in which all nodes are distributed randomly. The mobile aent mappin with link probability detection in findin the optimal stable path between the nodes connectin the ateway is the first approach of this kind Route Maintenance The second stae in the proposed protocol is route maintenance in local MANET and to Internet. In local MANET it is responsible to keep paths to local destination by ettin special packets i.e. hello messaes from intermediate nodes. In Internet it is responsible to keep paths to the ateway throuh intermediate nodes for hih network connectivity. The route maintenance to ateway for ettin services from Internet is done automatically usin the data packets forwarded throuh the intermediate active nodes selected in ateway discovery. The mobile aents traversin the link establish a pheromone track by settin the pheromone values from the source node to the ateway node. The path established keeps updatin the pheromone value in association with the decay factor, i.e., pheromone decay time. The pheromone decay time is set based on the link life time predicated by mobile aents. The pheromone value ets increased by a predefined value to maintain the route based on the sensitivity factor F for each data packet traversed. The pheromone value ets decreased by a predefined value to know the path is not used any more. When a node v relays a data packet toward the i ateway v to a neihbour node v, it increases the pheromone value of the entry ( v, v, ϕ) by ϕ, i.e., the path to the ateway is strenthened by the data packets. In contrast, the next hop v increases the pheromone value of entry ( vs, vi, ϕ) by ϕ, i.e., the path to the source also strenthened. The evaporation process of the real pheromone is simulated by reular decreasin of the pheromone value accordin to the Equation 4 iven below. ϕ = (1 q). ϕ q [0,1] (4) The mobile nodes reconize duplicate reception of data packets, based on the source address and sequence number in IPv 6 routin header. If a mobile node receives a duplicate packet, it sets the DUPLICATE_ERROR fla and sends the packet back to the previous node. The previous node deactivates the link to this node, so that data packets are not sent in this direction any more Route Failure Handlin The third stae in the proposed protocol is route failure, due to the active intermediate or source nodes havin mobility which chanes the ateway access. The proposed protocol reconizes the route failure in local MANET throuh a missin acknowledment from MAC layer. If a mobile node ets a ROUTE ERROR messae for a certain link in local MANET, it first deactivates that link by settin the hop count to 0.

6 30 The International Arab Journal of Information Technoloy, Vol. 5, No. 1, January 2008 The new route discovery process started usin AODV messaes. For Internet the routin table entries in the intermediate or source node has a pheromone decay time predicated by mobile aents durin ateway discovery. The pheromone decay time ets decreased until it reaches a threshold value. The threshold value K ives an indication that link break is oin to occur in the path to the ateway. The mobile aents are broadcasted to rediscover a new route or to update the old route to the ateway. If the mobile active node in the path to the ateway is movin it sends a mobile aent with the source address available in the routin table. The mobile aent searches for an alternate link to the ateway. If a second ateway link available in the routin table with pheromone value and pheromone decay time not reachin the minimum threshold value then data packet sent in this path. Either the data packet can be transported to the ateway or to local active MANET node else the backtrackin continues to the source node. 4. Performance Evaluation The performance evaluation of proposed protocol compared with existin proposals of Wakikawa and Jeler usin the network simulator ns-2. The source code of AODV protocol modified with the proposal of Wakikawa as AODV Wahiawa and OLSR protocol modified with the proposal of Jeler as OLSR Jeler. The proposed protocol AODV ALMA uses mobile aent code available as open source. The simulation scenario description is iven in Table 3 with the traffic type set for all nodes as same ftp traffic Simulation Results The main performance metrics under consideration are those recommended in RFC 2501: Packet Delivery Ratio (PDR): the percentae of successfully delivered data packets over all the data packets sent out by the sources. Normalized control overhead: the total number of control packets sent and forwarded over the number of successfully delivered data packets. This metric ives the cost in terms of the overhead in different approaches. Gateway discovery overhead: the total number of control messaes associated with the discovery of ateway. This metric ives information about the control overhead to provide Internet connectivity. Connectivity: the averae number of mobile nodes required in hybrid MANET to et unexpired routes. End-to-End delay: time taken to send data packets from source to destination normalized with the time taken to discover the route Packet Delivery Ratio (PDR) Fiure 3 ives the comparisons results of PDR for AODV ALMA, AODV Wakikawa, and OLSR Jeler. Table 3. Simulation scenario. Transmission rane 250 mts Physical layer two-ray propaation Link layer IEEE Queue buffer size 64 packets Routin buffer size 50 packets Channel capacity 2 Mbps No. of ateway 3 No. of mobile node 50 Area 1500 x 300 mts Speed 0-10 m/s Pause time sec Traffic type CBR Sources 10 Sendin rate 4 packets/sec Packet size 512 bytes Mobile aent packet 10 History size 6 hops Simulation time 900 seconds The PDR of AODV ALMA is ivin hih throuhput with that of the reactive AODV Wakikawa. The chane of performance for hih pause times caused when the network is quite sparse and static ivin a condition that nodes become unreachable for lon periods. The OLSR Jeler has very less PDR compared to the proposed protocol. The reason for this in OLSR Jeler is that link break in selected path to either ateway or local MANET handled after some time ap. In AODV ALMA the link break to the ateway is monitored by the chane in the pheromone value. The pheromone decay time associated with life time of the link. The route maintenance is done with the threshold value for each link in the path to the ateway. After carefully analyzin the simulation results, AODV ALMA use the pheromone value to adaptively select next hop to the ateway ivin topoloy chanes stron to mobility. Fiure 3. Packet delivery ratio.

7 Internet Connectivity for Mobile Ad Hoc Networks Usin Hybrid Adaptive Mobile Aent Protocol Normalized Control Overhead Fiure 4 ives the overhead comparison results for AODV ALMA, AODV Wakikawa, and OLSR Jeler. The OLSR_Jeler due to its proactive nature has hih control overhead in hih mobility conditions. As the mobility become static overhead decreases but not as less than AODV Wakikawa. The AODV Wakikawa due to its reactive nature has less control overhead in hih mobility conditions and reducin further as the network becomes static. The proposed AODV ALMA has a control overhead that is optimized to that of AODV Wakikawa in hih mobility conditions and 4.5. Connectivity Maintenance The connectivity of nodes ives path for data packets to be forwarded without loss and comparison results shown in Fiure 6. The AODV ALMA achieves hih connectivity in hih mobility conditions due to its adaptive pheromone value chanes and updatin by data packets forwarded to the ateway. The connectivity in the path tracked by link probability values predicated by mobile aents selected for forwardin. In OLSR Jeler connectivity is slihtly lesser than AODV ALMA with proactive route updatin. In AODV Wakikawa connectivity is very less due to the rediscovery of new routes to the ateway and to chanes in link break for hih mobility conditions. slihtly hiher for low mobility conditions. Fiure 4. Normalized control overhead Gateway Discovery Overhead The ateway discovery functions different in all the proposals and comparison results shown in Fiure 5. The AODV ALMA ateway discovery approach overhead is less in hih and low mobility conditions. The reason is the size and number of mobile aent packets multicast with priority to discover the path to destination simultaneously with local MANET messaes. The AODV Wakikawa approach ateway discovery overhead is hiher in hih mobility conditions. The reason is that time value (TTL) set to discover the ateway does not chane for mobility conditions. The OLSR Jeler due to its limited periodic floodin approach the ateway discovery overhead is hiher compared to that of AODV ALAM and varies with mobility. Fiure 6. Connectivity Maintenance End-to-End Delay The end-to-end delays include bufferin, queuin delay at interface queue, retransmission delay, propaation and transfer time delay. The assumption for simulation is that all other delay values are same except the route discovery delay. Fiure 7 ives the comparison results of end-to-end delay. Fiure 7. End-to-End Delay. Fiure 5. Gateway discovery overhead. The proposed protocol AODV ALMA has less delay with separate discovery process simultaneously for ateway and for local MANET. The different mobility condition ives the end-to-end delay variable due to chanes in pheromone value. The OLSR Jeler proactive approach discovers route to destination periodically ivin slihtly hiher delay due to the time ap in findin the route separately. The AODV

8 32 The International Arab Journal of Information Technoloy, Vol. 5, No. 1, January 2008 Wakikawa reactive approach discovers route to destination when needed by first searchin in local MANET and then to ateway has hiher delay. The time extension value (TTL) for initial route discovery in MANET up to network diameter and after some time for the ateway vary with mobility conditions. 5. Conclusion This paper proposed a new hybrid adaptive ateway discovery mechanism for hybrid MANET usin mobile aents combined with reactive distance vector. The performance tradeoffs and limitations with existin proposals are compared with the proposed protocol usin simulation. The proposed protocol ives hih network connectivity to ateway and reduced end-toend delay in packet transmission to destination. The results compared for mobility condition only, it can also be tested for different traffic and scalability conditions. The proposed protocol can be extended with other internetworkin related issues such as improved DAD mechanism, efficient support of DNS, discovery of application and network services, network authentication and interated security mechanism. References [1] Ammari H. and El-Rewini H., Performance Evaluation of Hybrid Environments with Mobile Gateways, in Proceedins of the 9th International Sympesium on Computin and Communications, vol. 1, no. 1, pp , June [2] Broch J., Maltz D., and Johnson D., Supportin Hierarchy and Heteroeneous Interfaces in Multi- Hop Wireless Ad Hoc Networks, in Proceedins of the Workshop on Mobile Computin, Australia, pp , June [3] Choudhary R., Bhandhopadhyay S., and Paul K., A Distributed Mechanism for Topoloy Discovery in Ad Hoc Wireless Networks Usin Mobile Aents, in Proceedins of Mobicom, pp , Auust [4] Jeler C., Noel T., and Frey A., Gateway and Address Auto Confiuration for IPv6 Ad Hoc Networks, Internet-draft, April [5] Lee J., Hybrid Gateway Advertisement Scheme for Connectin Mobile Ad Hoc Networks to the Internet, in Proceedins of the 57th IEEE VTC 2003, vol. 1, no. 1, pp , April [6] Lei H. and Perkins C., Ad Hoc Networkin with Mobile IP, in Proceedins of the Second European Personal Mobile Communications Conference, Germany, pp , [7] Matsuo H. and Mori K., Accelerated Ants Routin in Dynamic Networks, in Proceedins of the International Conference on Software Enineerin, Artificial Intellience, Networkin and Parallel /Distributed Computin, pp , [8] Ratanchandani P. and Kravets R., A Hybrid Approach to Internet Connectivity for Mobile Ad Hoc Networks, in Proceedins of IEEE Wireless Communications and Networkin Conference, pp , March [9] Ruiz P. and Gomez-Skarmeta A., Adaptive Gateway Discovery Mechanisms to Enhance Internet Connectivity for Mobile Ad Hoc Networks, Ad Hoc and Sensor Wireless Networks, vol. 1, pp , March [10] Scott M., Joseph P., and Greory H., Internet Based Mobile Ad Hoc Networkin, IEEE Internet Computin, vol. 3, no. 4, pp , July-Auust [11] Sinh S., Perkins C., Ruiz P., and Clausen T., Ad Hoc Network Auto Confiuration: Definition and Problem Statement, 00.txt, July [12] Wakikawa R., Malinen J., Perkins C., Nilsson A., and Tuominen A., Global Connectivity for IPv6 Mobile Ad Hoc Networks, (work in proress), fts/draft-wakikawa-manet-lobalv6-02.txt, October Velmuruan Ayyadurai is a lecturer in electronics enineerin department in Madras Institute of Technoloy at Anna University, India. He holds MSc deree in microwave and optical enineerin from Madurai Kamara University in He is currently pursuin his research work for PhD deree in information and communication enineerin. He has more than 10 years of experience in teachin and research. He is supervisin MSc student thesis in wireless internet, wireless security and wireless TCP/IP interation. He teaches courses at both BSc deree and MSc deree levels in electronics and communication enineerin. Raaram Ramasamy is a professor & head of the Department of Computer Science & Enineerin, Thiaaraar Collee of Enineerin, Madurai. He has a BSc Deree in electrical and electronics enineerin from University of Madras in He secured the MSc in electrical enineerin from IIT Kharapur in 1971 and the PhD deree from Madurai Kamara University in He

9 Internet Connectivity for Mobile Ad Hoc Networks Usin Hybrid Adaptive Mobile Aent Protocol 33 has been teachin computer hardware and software and supervisin research activities. He has published nearly 32 papers in ournals, seminars and symposia. His areas of interest are neural networks, fuzzy systems, and enetic alorithms. He has published more than 11 text books for schools and collees in Enlish and one book in computer science in Tamil. He served at the Makerere University, Kampala, Uanda durin and at the Mosul University, Iraq durin He secured two best technical paper awards from the Institution of Enineers (India), one from the Indian Society of Technical Education.