Inter national J our nal of Electrical and Computer Engineering (IJECE) V ol. 6, No. 1, February 2016, pp. 421 430 ISSN: 2088-8708 421       I ns t it u t e  o f  A d v a nce d  Eng ine e r i ng  a nd  S cie nce   w     w     w       i                       l       c       m     Issues in Routing Mechanism f or P ack et F orwarding : A Sur v ey Rohit Nilkanth De vikar * , Dipak V . P atil ** , and V . Chandraprakash * * Department of Computer Science and Engineering, K L Uni v ersity , Green Fields, V addesw aram, Guntur , (A.P .) 522502, India ** Department of Computer Engineering, Gokhale Education Society’ s R. H. Sapat Colle ge of Engineering, Management Studies and Research, Nashik - 422 005, (M.S.), India Article Inf o Article history: Recei v ed May 18, 2015 Re vised Sep 8, 2015 Accepted Oct 2, 2015 K eyw ord: Routing protocols P ack et forw arding Connecti vity Load balancing Con v er gence Congestion Control ABSTRA CT No w adays internet has become more popular to each and e v ery one. It is v ery sensiti v e to nodes or links f ailure due to man y kno wn or unkno wn issues in the netw ork connecti vity . Routing is the important concept in wired and wireless netw ork for pack et transmission. During the pack et transmission man y times some of the problems occur , due to this pack ets are being lost or nodes not able to transmit the pack ets to the specific destination. This paper discusses v arious issues and approaches related to the routing mechanism. In this paper , we present a re vie w and comparison of dif ferent routing algorithms and protocols proposed recently in order to address v arious issues. The main purpose of this study is to address issues for pack et forw arding lik e netw ork control management, load balancing, congestion control, con v er gence time and instability . W e also focus on the i mpact of these issues on pack et forw arding. Copyright c 2016 Institute of Advanced Engineering and Science . All rights r eserved. Corresponding A uthor: Rohit Nilkanth De vikar Research Scholar Departement of Computer Science and Engineering K L Uni v ersity , V addeshw aram, Guntur +919028339491 Email: rohit.de vikar89@gmail.com 1. INTR ODUCTION In the old era, routing is simply forw arding the pack et from one node to another , b ut no w it is the process of choosing the best optimal path for pack ets transmission to impro v e the netw ork performance. Basically the telephone switching netw ork pro vides the quality of service by establishing the connection between sender and recei v er before transmission of data. In telephone switching netw ork delays are introduced during connection establishment phase, connection releas e phase as well as in transmission of pack ets. The performance of data transmission in telephone netw ork is good b ut it requires a bandwidth dedication between sender and recei v er . Due to this the utilization of netw ork links is poor , results in reducing the performance of the o v erall netw ork. Instead the pack et switching of fers a best ef fort deli v ery of the pack ets. P ack et switched netw ork does not require the conn e ction establishment between sender and recei v er . So, delay is only present at the time of data transmission. During the transmission of the pack ets in wired as well as in wireless netw orks man y of the problems stand in front of us which will be discussed in the ne xt section. Some important approaches and issues considered in this paper re g arding routing techniques are: 1. Control management related issues present in wired as well as in wireless netw ork. 2. Dif ferent types of congestion controlling techniques for T raf fic analysis. 3. Netw ork load balancing mechanism in T raf fic Engineering. 4. Problem occurring during scaling of netw ork. Also the ef fect of scaling on routing table. 5. Choosing the better techniques to impro v e the netw ork a v ailability . 6. Instability of the netw ork is the current issue in the routing mechanism which results in loss of pack ets due to fluctuation of the netw ork links called as route flapping. 7. Dif ferent algorithms on im pro ving the con v er gence time, so that the net w o r k updates with minimum tim e duration. J ournal Homepage: http://iaesjournal.com/online/inde x.php/IJECE       I ns t it u t e  o f  A d v a nce d  Eng ine e r i ng  a nd  S cie nce   w     w     w       i                       l       c       m     DOI:  10.11591/ijece.v6i1.8151 Evaluation Warning : The document was created with Spire.PDF for Python.
422 ISSN: 2088-8708 W e will discuss the abo v e issues one by one: 2. R OUTING CONTR OL MAN A GEMENT Ov er the years, wired and wireless netw ork security has become a major issue. Netw ork pro vides security with authentication, authorization, denial of services, IP security . One time passw ord (O TP) is v ery ef fecti v e security mechanism no w adays. The entire go v ernment sector mak es the use of O TP for better security . Madalina Baltatu et al [1] focused on todays internet, as it uses the TCP/IP for communication, e v en though problem may occur at the time of authentication. The y describe the att acks using ICMP messages. The ICMP mes- sages generally consist of Destination unreachable and T imeout e xceeded messages. The Denial of service attack e xploits one of these tw o messages. The author describes the security attacks using ICMP router disco v ery messages. The y use the protocol called ICMP router disco v ery protocol (IRDP) [1]. During the adv ertisement the IRDP might result in f acing the follo wing attacks: passi v e monitoring, man-in-the-middle, denial of service etc. The most secure protection for routing is the implementation of static route from source to destination. But it of fers for only small netw ork or in LAN, with no an y special QoS requirements. But for dynamic traf fic flo w which requires QoS static routing f ails. In this case there is a need to use suitable, reliable routing protocols which pro vides QoS and authentica- tion mechanism is mandatory . RIP and OSPF are firstly tak en into consideration for pro viding QoS and authentication. Geof f Huston et al [2] pro vide the security related issues in BGP . BGP consists of speak er node which cont ains the path information of ASs. F o r AS 100 speak er node is C and for AS 200 only single router acts as speak er node. BGP does not pro vide protection ag ainst replay , message insertion, message deletion, man-in-the-middle attack, message modification type of attacks. The y are combined with TCP to pro vide protection ag ainst all of the abo v e attacks. The secure BGP (S-BGP) addresses this vulnerability . S-BGP pro vides the three security mechanism [3]. First it uses Public K e y Infrastructure (PKI) for authentication of o wnership of IP address blocks. Second BGP path attrib ute is used to carry the digital signatures. Third, IPSec is used to pro vide the data and partial sequence inte grity and to enable the BGP speak er nodes to authenticate each other [4]. Figure 1. Communication between autonomous systems Dan W endlandt et al. [5], [3] use A v ailability Centric Routing (A CR) which enable end system to pro vide secure communication e v en if adv ersary controls the netw ork infrastructure. A CR uses the four components: 1) It uses transit Autonomous system (AS) which pro vides multiple routes for each destination. Due to t his adv ersary f ails to track a v alid communication path. 2) The destination can be identified by cryptographic algorithm by source A C R to conform whether des tination is v alid. 3) End systems securely monitor s the communication to measure the performance, if performance is lo w then it chooses another path. 4) A CR end systems distrib ute traf fics o v er more than one path. The A CR pro vides end to end security with increasing the a v ailability of the netw ork [5]. Chin-Fu K uo et al [6] proposes the dynamic routing algorit hm that could select the path randomly for pro vid- ing better security . The main goal is to propose the distance v ector based algorithm to impro v e the data transmission for dynamic routing. The tradition routing table for distance v ector routing us es the routing information protocol (RIP) consist of field (Ne xt Hop, Destination Address, Metrics). The author uses the distrib uted dynamic routing algorithm (DDRA) which is the e xtended routing table v ersion of distance v ector routing algorithm. The table k eeps the entries same as RIP with added field of History Record for P ack et Deli v eries to the De stination Node [6]. Mael Saleh and Liang Dong [7] proposed a security enhancement in pack et switched netw ork for real time application. The IJECE V ol. 6, No. 1, February 2016: 421 430 Evaluation Warning : The document was created with Spire.PDF for Python.
IJECE ISSN: 2088-8708 423 T able 1. Security Related T echniques Article T echnique for Secure Com- munication Protocol/ Algorithm Adv antages Madalina Baltatu et al [1] Uses ICMP router Disco v- ery messages IRDP It pro vides the route adv ertisement and route solicitation. Glenn Jacobson [3] Uses PKI, BGP P ath At- trib utes, IPSec S-BGP It addresses the vulnerability of au- thentication and authorization. Dan W endlandt et al. [5] Uses T ransit AS, Crypto- graphic Algorithm, End sys- tem Monitors the communi- cation, A CR Pro vides end to end security . Chin-Fu K uo et al [6] Random pack et deli v ery , Maintain External Routing Information DDRA Adding a field in routing table of RIP History Record for P ack et De- li v eries to the Destination Node Mael Saleh and Liang Dong [7] A single layer Module, Mul- tilayer module. Dif f-EDF schedule with security en- hancement services Use for Real T ime Application use of adapti v e security a w are scheduling for pack et switched netw orks has been done in [1] which is incorporated with security enhancement services. F or scheduling the y use the dif ferentiat ed-earliest-deadline-first (Dif f-EDF) and for security me chanisms the y use security enhancement services [7]. The tw o modules are used to design a security service enhancement. 1) A single-layer Module, 2) Multilayer module. The single la y e r design pro vides the enhancement for security services lik e confidentiality , inte grity , or au- thentication. Whereas the multilayer module pro vides the enhancement for multilayer security services. At the des- tination the system measures the use of b uf fer and informs the source to change the security le v el of data pack ets adapti v ely . This scheme is ef fecti v e for time and security related application. T able 1 pro vides the comparison of dif ferent security related issues for controlling the netw ork. 3. CONGESTION CONTR OL TRAFFIC AN AL YSIS In data transmission and queuing theory , congestion occurs when the link carries more traf fic than its capacity . The congestion in the netw ork af fects on pack et transmission, data loss, as well as more queuing dela y [8]. In the ne w era, the internet users are gro wing v ery rapidly . Due to this man y times congestion may occur during the pack et transmission. The congestion is the main reason to de grade the performance and quality of service of the netw ork. The follo wing section describes the v arious issues and controlling techniques for congestion in the netw ork: Sa v erio Mascolo [8] proposed the control la w and smith principle for managing the ABR (A v ailable Bit Rate) flo w in A TM netw ork for congestion control. The Laplace transform technique is us ed to design the controller and for analyzing the performance of control system. F or each flo ws from the switch the queue is maintained that stores the pack et for transmission. Ian F . Ak yildiz et al. [9] discussed about the congestion control scheme for satellite netw ork. T raditional TCP approach has the lo wer throughput mainly because of lar ge propag ation delays and high link error rate. The author introduces TCP -PEA CH control schem e which pro vides end to end solution to impro v e the throughput performance. The TCP-PEA CH control scheme has follo wing algorithms: 1) Sudden start 2) Congestion A v oidance 3) F ast retransmission 4) Rapid reco v ery The congestion a v oidance and f ast retransmission are the basic TCP-RENO and TCP-VEGO algorithms [9]. The sudden start and rapid reco v ery w ork on the principle of dummy se gments. The dummy se gments are the lo w priority se gments generated at the sender . These dummy se gments are used to a v oid the congestion in the netw ork. Generally the sender sends the dummy se gments to e xamine the a v ailability of the netw ork resources. If the router Issues in Routing Mec hanism for P ac k et F orwar ding : A Surve y (R. N. De vikar) Evaluation Warning : The document was created with Spire.PDF for Python.
424 ISSN: 2088-8708 T able 2. Controlling Congestion in the Netw ork Article Netw ork Algorithm for Congestion Con- trol Description Sa v erio Mascolo [8] A TM Control Flo w and Smiths Prin- ciple Use to control best ef fort traf fic, algo- rithm guarantees stability of netw ork queue also pro vides full link utilization. Ian F . Ak yildiz et al. [9] Satellite TCP-PEA CH Impro v es the f airness for sharing net- w ork resources and goodput perfor - mance. Atilla Eryilmaz et al. [10] W ireless primary-dual congestion con- trolling (PDCC) Pro vides f airness and stability in wire- less netw ork. Jianhua He et al. [11] V ehicular Cross layer design approach and traf fic rate control approach These control schemes are used to adapt dynamic traf fic load ef fecti v ely . Haitao W u et al [12] Data Centre Incast congestion control for TCP (ICTCP) A v oid Congestion by Achie ving zero timeout and pro vides better goodput. Shikhar Shukla et al. [14] Data Centre PLA T O TCP Use pack et labeling technique for de- tection of loss pack ets. Ferhat Dikbiyik et al. [15] OpticalWDM Prepro visioning, Backup repro- visioning, Hold-lightpath Increases a v ailability of connection and decreases connection setup time. in the path is congested then it discards the dummy se gment and, if not then router replies to the sender with A CK message to inform the sender that i am ready to recei v e the pack ets. Atilla Eryilmaz et al. [10] reduce one of the f air rate allocations of resources problems with the use of primary-dual congestion controlling (PDCC) mechanism incorporated with backpressure. The author proposes the mathematical model to impro v e f airness and stability . 3.1. V ehicular Netw ork The Jianhua He et al. [11] proposed dedicated short range communication (DSRC) based collaborati v e safety application (CSA) scheme for congestion control. It contains tw o adapti v e congestion control mechanism. 1) Detec- tion of congested channel has been done at MA C layer . 2) After detection the congestion signal is transmitted to the application layer to control the traf fic rate. Haitao W u et al [12] propose the ne w technique called Incast congestion control for TCP (ICTCP). Actually what happens in traditional approach the multiple users send data to the single user or node or link, so there is a possibility of congestion at the recei v er . Unlik e in ICTCP the y focused on rec ei v er based congestion control algorithm to reduce the pack et losses. The [12] [13] pro vide the congestion control in pack et transmission, b ut the y f ail to reduce the detection of pack et losses at required e xtent. The Shikhar Shukla et al. [14] proposed the PLA T O TCP congestion control mechanism which detects the pack et losses. PLA T O concerns the pack et labeling scheme for the solution to TCP Incast. TCP detects the loss after 200ms, because of loss cannot be a v oided earlier . The PLA T O detects the loss pack ets wi th the help of pro viding three duplicate ackno wledgement mechanisms with labeling system rather than w aiting for 200 ms delay . Ferhat Dikbiyik et al. [15] proposed the problem in e xploiting the e xcess capacity (EC) in optical WDM (W a v elength Di vision Multiple xing) in terms of bandwidth blocking. Optical netw ork uses the tw o links: primary link and protected link (Backup link). The primary link is used to transmit re gular traf fic and when an y f ailure or congestion occurs in primary link or in an y node then there is a pro vision to use the protected link. Congestion controlling mechanism is v ery important part in traf fic engineering, as it leads to the birth of load balancing. F or congestion controlling it is necessary to di vide the traf fic across multiple paths or store it at the nodes (router or switch) queue. T able 2 pro vides dif ferent congestion control mechanism for pack et forw arding. IJECE V ol. 6, No. 1, February 2016: 421 430 Evaluation Warning : The document was created with Spire.PDF for Python.
IJECE ISSN: 2088-8708 425 4. NETW ORK LO AD B ALANCING Load balancing is a technique used to distrib ute the w orkload o v er multiple paths or multiple processors. Load balancing term is generally e v olv ed to utilize the netw ork resources ef fecti v ely and reduce the congestion in the netw ork. General Load balancing algorithms described in [16] are weighted balance, Priority , Ov erflo w , Persistence, Least Used, Lo west Latenc y . The basic load balancing algorithm to impro v e the netw ork a v ailability is sho wn in figure 2. In the netw ork for switching of traf fic a primary paths as well as number of alternati v e paths are present. In traditional approach there is a use of primary path for the transmission of traf fic and when the primary path is congested or f ails, then and then only the secondary path is used. The alternate or backup route is utilized only when the primary link f ails which sho ws inef ficient uti lization of bandwidth and netw ork resources. The simple algorithm sho wn belo w describes ho w the netw ork balances the load on the primary as well as on secondary paths. Suppose P1 and P2 represent P ath1 and P ath 2 respecti v ely . S is the source node and D is the destination node. The total number of routers between source and destination is R. The number of routers in path1 and path2 are assumed to be R1and R2 respecti v ely . N is the total number of pack ets at the source at time T0. Figure 2. Algorithm and relati v e figure for Load balancing approach Suppose source S has 90 pack ets at time T0, and then it will send 60 pack ets from path2 and 30 pack ets from path1. Xiaohua Jia et al [17] analyses the shortest path tree (SPT) and Minimum spanning tree (MST) approach to reduce the delay and cost of the netw ork respecti v ely . Also the MST and SPT are used to pro vide the optimum solution for load balancing and w a v elength assignment problem. Kartik Gopalan et al [18] proposed a Link criticality based routing (LCBR) algorithm used to select the primary route and primary backup route. In the netw ork Dijikstra algorithm is used to select the shortest path from source to destination b ut it is not ef ficient in the presence of lar ge delay also Dijikstra algorithms cannot s olv e the end to end delay partitioning problem. The route selection process in Primary LCBR can be carried out using both online and of fline f ashion. In of fline technique performs once for entire netw ork. Between each source and destination the k shortest paths are calculated and compute the e xpected load on each link. While in online technique the Primary LCBR calculates cost of each link between source and destination. It also checks the resources a v ailable for the route r satisfy the QoS. If resources are suf ficient then PLCBR distrib utes end to end delay among the links of route r . After this PLCBR calculates the remaining link capacity R of route r and projected netw ork cost cost(G). There are tw o conditions to reject the route setup request for primary path F . 1) If route does not ha v e resources to fulfill the QoS, 2) The v alue of cost(G) for route r is greater than predefined cost. F or selection of Primary Backup LCBR (PBLCBR) the primary route elements i s remo v ed and same procedure lik e PLCBR is applied to obtain the Backup route [18]. Lu Ruan et al. [19] describe the load balancing in optical WDM netw ork based on primary , backup and free channels. The primary link can b e selected as a shortes t path with more number of free channels. A. K. Mishra and A. Sahoo [20] proposed the S -OSPF algorithm for load balancing. The S-OSPF w orks based on traf fic demand which maintains the traf fic matrix. But it requires more processing po wer and the fluctuation in the netw ork will increase the problem of maintaining the traf fic matrix. Jiayue He and Jennifer Re xford in [21] proposed the method for multipath routing mechanism for load balancing. The y e xplain the congestion in control plane and data plane and splitting the traf fic o v er multiple paths. The router in the data plane use to forw ard the pack ets on the alternate path based on encapsulation and e xplicit routing techniques. Marija Anti et al. [22] proposed linear programming approach for load balancing. The y define t he link load not on l y depends on node traf fic load, traf fic matrix element. The Sangsu Jung et al. i n [23] propose the load balancing and congestion cont rol approach in Issues in Routing Mec hanism for P ac k et F orwar ding : A Surve y (R. N. De vikar) Evaluation Warning : The document was created with Spire.PDF for Python.
426 ISSN: 2088-8708 wireless mesh netw ork. The traditional routing approach such as A OD V and geographic routing such as GPSR are responsible for long delay and frequent pack et losses if the hot spot is congested. The wireless mesh netw ork pro vides the hub and spok e type routing polic y in which first ly the y reflect the traf fic v olume, secondly the finite element method is used to assign a potential v alue to each node, and finally proposed a no v el Potential-Field Based Routing (PFBR) protocol for load balancing. The potential v alue reflects distance to destination as well as traf fic v olume at each node. In PFBR the congestion at f ar node can be a v oided by only e xchanging the local routing information with the neighbor which is just a one hop distance from the node. The mesh node can forw ard a pack et to the neighbor node which has lo wer potential v alue [23]. Marija Antic et al. [24] proposed the load balancing shortest path routing (LB-SPR) technique, which signif- icantly reduce the reliable netw ork cost. The LB-SPR pro vides: 1) It does not require the actual traf fic pattern for routing optimization. 2) Require lo w cost for bandwidth reserv ation. 3) Ev ery router in the netw ork kno ws the maximum traf fic load it is allo wed to generate. The LB-SPR algorithm impro v es the reliability of the netw ork. If an y node or link f ails the LB-SPR with OSPF adjust the routing in such a w ay that traf fic loads are forw arded. LB-SPR consists of tw o steps to send the traf fic from source to destination. In this algorithm the source does not directly transmit the traf fic to destination. It first sends the traf fic to the intermediate node called load balancing router , after that from load bal ancing router to the destination [22], [24]. T o reduce the problem of traf fic demand [20], e xt ended scheme of S-OSPF is used in [25], which w orks on hose model. In this scheme only the total amount of traf fic the source sends into the netw ork and the total amount of traf fic it recei v es from the netw ork is needed. Suppose the netw ork consists of source “p” and destination “q” . The source “p” sends the traf fic into the netw ork (Outgoing traf fic) is gi v en by the follo wing equation: X q 2 Q d ( pq ) p p 2 Q (1) Where, p is the maximum traf fic sent by source into the netw ork. While the total incoming traf fic to the source node “p” is gi v en by: X p 2 Q d ( pq ) q q 2 Q (2) Where, q is the maximum traf fic recei v ed by node “q” from the netw ork. The Brice Augustin in [26] proposed the multipath routing in the netw ork with three dif ferent load balancing algorithms: Per -flo w , Per -Destination, and Per -P ack et. The y est imate the results using these algorithms and observ e that the per -pack et load balancing are less ef fecti v e. In per -pack et load bal ancing the main problem is each pack et follo ws dif ferent paths so it requires the reordering of the pack ets. While the per -flo w and per -destination load bal- ancing techniques are still widely used. T raf fic engineering (TE) plays an important rol e in determining the reliability and performance of a netw ork. Geor ge Athanasiou et al. [27] describe the Ener gy A w are T raf fic engineering scheme. The main objecti v e of TE is to minimize the maximum utilization of link. The author also focuses on minimizing the ener gy consumption by turning the unutilized and idle links into the sleeping mode. Load balancing can be achie v ed by splitting the traf fic o v er multiple links with consideration of minimizing the maximum link utilization. Equal cost multiple paths forw arding (ECMP) is a v ery prominent technique for the netw ork in which each link has equal cost. But this technique is inef ficient for increasing path di v ersity in which each link is ha ving dif ferent weights to optimize netw ork resource usage. In [28] uses the PEFT (P anelized Exponential Flo w spliTting) technique which transmits the pack ets o v er multiple unequal link cost paths, where as traf fic splitting decision are made independently . The topology uses the OSPF protocol for computing the links weight. F or implementation of link weight optimizer requires solving a con v e x optimization problem and Link weights computation. T raf fic enginee ring is impro v ed by f airer netw ork load balancing, minimizing the maximum link utilization (MLU) and increasing netw ork capacity . Shuo F ang et al. [29] proposed a softw are defined a p pr o a ch for load balancing by inte grating dynamic load balancing multipath scheme (DLBMP) with congestion control (CC). The DLBMP+CC impro v e the netw ork throughput by making the full uti- lization of bandwidth and congestion control used to pre v ent the e xcessi v e n e tw ork traf fic entering into the netw ork. W ith the use of Softw are Defined Netw orking (SDN) in [29], the dynamic algorithm can react v ery quickly to w ards topology changes, congestion control, load imbalance, and an y updates. F or SDN, W idhi Y ah ya et al. in [30] proposed a load balancing algorithmic approach based on e xtended Dijkstra algorithm. This algorithm is used to find the nearest IJECE V ol. 6, No. 1, February 2016: 421 430 Evaluation Warning : The document was created with Spire.PDF for Python.
IJECE ISSN: 2088-8708 427 T able 3. Netw ork Load Balancing Article Load Balancing Scheme Description Xiaohua Jia et al [17] SPT , MST SPT is used to minimize Delay and MST reduces the net- w ork Cost. Kartik Gopalan et al [18] LCBR Pro vides higher resource utilization, also ha v e primary as well as backup route for f aster link f ailure reco v ery . Lu Ruan et al. [19] Routing with Load Balanc- ing Heuristics (RLBH) Assign link cost based on number of free channels a v ail- able at link. A. K. Mishra et al. [20] Smart-OSPF (S-OSPF) Pro vides better resources utilization for traf fic engineer - ing. Jiayue He et al. [21] Fle xible Multipath Routing Fle xible spli tting Distrib utes load o v er multiple paths based on traf fic class. Marija Anti et al. [22] Linear Programming Ap- proach Assign traf fic v alues to netw ork nodes. Sangsu Jung et al. in [23] Potential-Field Based Rout- ing (PFBR) protocol Assign potential v alues to nodes based on numerical anal- ysis. This potential v alue pro vides distance to destination as well as traf fic v olume. Marija Antic et al. [24] LB-SPR Supports guaranteed traf fic load and simplifies resource reserv ation. Brice Augustin [26] Per -flo w , Per -Destination, and Per -P ack et Pro vides broad description of multipath routing in the in- ternet. Geor ge Athana- siou et al. [27] Ener gy A w are T raf fic engi- neering Splitting the traf fic o v er multiple links by minimizing the maximum link utilization. Fung Po Tso et al. [28] PEFT (P anelized Exponen- tial Flo w spliTting) Use to transmit pack ets o v er unequal cost path. Shuo F ang et al. [29] DLBMP+CC Only P ath load attrib ute is monitor to track netw ork traf fic load. serv er by considering edge weights as well as node weights. T able 3. sho ws dif ferent scenarios for balancing the load o v er netw ork. 5. INST ABILITY AND CONVERGENCE TIME In the recent years internet instability and route fluctuation are important problems in the netw ork. Instability in the netw ork results in loss of pack ets, which in turn increases the latenc y and con v er gence time. F ollo wing section describes the research w ork done by dif ferent researchers to impro v e the con v er gence time and instability of the netw ork. Craig Labo vitz et al. [31] describe some unpredicted trends in routing stability . The authors de v eloped the taxonomy for routing information and identify the origin of pathological beha vior . The routing information is classified into three classes: 1) F orw arding i nstability , 2) Pol ic y Fluctuation, 3) P athologic updates (redundant). This research observ ed redundant data during update of routing topology for nine months. Most of the data collected were redundant. Finally the y e xplain the impact of this redundant data on netw ork infrastructure. Aristotelis Tsirigos et al. [32] define a f ailure probability technique to impro v e the stability of the netw ork. In mobile netw ork, the topology changes due to time unstable state of the netw ork. As this scheme defines the probability of path f ailure for each link, it is used to de v elop the probabili ty function Psucc (Probability that no more Mblocks is lost). From observ ation it w as sho wn that the probabi lity of successful communication increased between source and destination only when we increase the number of paths. This reduces the congestion and transmission delay . In the recent years OSPF and IS-IS are used to compute the shortest path tree (SPT) from router to router . As there are multiple SPT in the netw ork, reco v ery from f ailure causes changes in e xisting SPT topology which results in routing instability . P aolo Narvez et al. [33] [34] proposed the ne w algorithm which impro v es the stability of netw ork by making minimum changes in the e xisting SPT topology , when some li nk or router in the netw ork f ails. After f ailure, the discontinuity is encountered in link state adv ertisement and re-computation of routing table. F or link state routing OSPF (optimal shortest path first) protocol is used. As f ailures is increasing the instability in the netw ork increases. T o impro v e the f ailure resilienc y Issues in Routing Mec hanism for P ac k et F orwar ding : A Surve y (R. N. De vikar) Evaluation Warning : The document was created with Spire.PDF for Python.
428 ISSN: 2088-8708 without af fecting routing stability , Srihari Nelakuditi [35] proposed a f ailure insensiti v e routing (FIR) approach. This approach suppresses the link state adv ertisement. Using this approach, when at most one link f ailure notification is suppressed, a pack et is guaranteed to be transmitted to its destination along loop free path. The e xperimental results sho w that FIR pro vides better routing stability and a v ailability than OSPF in terms of netw ork sizes, f ailure frequenc y , and con v er gence delays. Y ang Richard Y ang et al. [36] reported the results on ef ficienc y and stability to achie v e the traf fic engineering objecti v es in interdomain routing when interactions among routing to multiple destinations cause instability in routing e v en if each route to destination has unique solution. Route selection problem i s stable only if the interaction among the ISPs follo ws the set of interdoma in traf fic engineering guidelines; otherwise instability occurs in route selection process. The accidental acti vities such as f ailure, misconfiguration, route flapping, induced se v eral BGP instabilities in the netw ork lead to delays, loss of data and connecti vity . T odays internet routers are o v ercome by a number of BGP updates caused by e v ents such as f ailure, session reset, and polic y changes. Such e v ents can delay routing con v er gence, which de grades the performance of netw orks in terms of jitter and delay sensiti v e application. W ei Sun et al. [37] propose the no v el approach of dif ferentiated processing in terms of BGP updates, which impro v e the routing con v er gence and reduces the routers load. Based on this approach the BGP updates are classified into tw o classes. Higher priority updates are processed sooner , while the lo wer priority updates are delayed to reduce router load and processing. Shi v ani deshpande et al. [38] proposed the BGP instability detection mechanism that can be e x ecuted by indi vidual router s. The input data for detection of instability is BGP update messages recei v ed by routers from its neighbor . From this BGP update messages features (lik e AS path length, AS path edit distance) are e xtracted in e v ery v e minutes, this sho ws the change in topology . The GLR (Generalized Lik elihood Ratio test), Se gmentation boundary detection, Boundary position optimization algorithms are used to detect the changes. Geof f Huston et al. [39] proposed a P at h Exploration Damping (PED) technique which reduces the v olume of BGP update messages and decreases the a v erage time required to restore reach- ability . The y compare PED impact on con v er gence time with Mean route adv ertisement interv al (MRAI), Route Flap Damping (RFD), and W ithdra w al Rate Limit ing (WRA TE). Mohammad Y anuar Hariya w an [40] compared dif ferent technoques lik e F ast Reroute one to one backs up, local rerouting, Haskin, 1+1 path protection reco v ery mechanism and PSL oriented path protection mechanism t echnique for f ast rerouting after f ailure. The performance sho ws that 1+1 path protection reco v ery mechanism has minimum pack et loss, b ut ha ving more cost. Rajvir Gill et al. [41] proposed the FLD-MRAI (Fle xible Load Dispersing MRAI) algorithm that disperses the load in the netw ork, which results in reducing the routers o v erhead. The authors focused on routing policies and their ef fects on number of updates, con v e r gence time. The FLD-M RAI algorithm w orks in case of both high and normal loads. When de gree of pref erence (DoP) chooses the shortest path, then FLD-MRAI belie v e this situation as normal load, and when DoP chooses the longest path then FLD-MRAI belie v e this situation as high load. Belo w table compares the dif ferent approaches to impro v e the con v er gence time of a netw ork. T able 4 pro vides dif ferent scenarios for impro ving the stability and reducing con v e gence time of netw ork. 6. CONCLUSION In this paper , we ha v e presented a brief surv e y on dif ferent issues lik e control management, a v ailability , congestion control, con v er gence time, instability , load balancing based on netw ork routing. Most of the researchers are w orking on netw ork load balancing and congestion control for traf fic engineering (TE). From the e xisting w ork we ha v e discussed the dif ferent problems and their respecti v e solutions for pack et forw arding by considering abo v e issues. I hope this paper will be beneficial to readers for better understanding about the current issues that are occurring in the netw ork for pack et forw arding. REFERENCES [1] Madalina Baltatu, Antonio Lio y , F abio Maino, Daniele Mazzocchi, Security Issues in Control, Management and Routing Protocols, T erena Netw orking Conference , May 22-25, 2000;1-12. [2] Geof f Huston, Mattia Rossi, and Gren ville Armitage, Securing BGP - A Literature Surv e y , IEEE Communications Surv e ys & T utorials , Second Quarter 2011;13(2): 199-222. [3] Glenn Jacobson , Security Issues with Internet Routing Border Gate w ay Protocol (BGP), Global Information Assurance Certification P aper , SANS Institute 2003. [4] Stephen T . K ent, Securing the Border Gate w ay Protocol: A Status Update, 7th IFIP-TC6 TC11 International Conference , CMS 2003, T orino, Italy , October 2-3, 2003; 40-53. [5] Dan W endlandt, Ioannis A vramopoulos, Da vid G. Andersen, Jennifer Re xford, Dont Secure Routing Protocols, Secure Data Deli v ery , In Proc. 5th A CM W orkshop on Hot T opics in Netw orks (Hotnets-V), (Irvine, CA) , No v . IJECE V ol. 6, No. 1, February 2016: 421 430 Evaluation Warning : The document was created with Spire.PDF for Python.
IJECE ISSN: 2088-8708 429 T able 4. Mechanism to Impro v e the Con v er gence T ime Article Protocol Description Aristotelis Tsiri- gos et al. [32] Multipath routing In this technique, routing scheme that uses multiple paths by splitting the information o v er multiple paths to in- crease corrected data recei v e probability . P aolo Narvez et al. [33] [34] Dynamic SPT Used to update only part of shortest path tree af fected by changes. Due to this only minimum changes ha v e been done to impro v e the stability . Srihari Nelakuditi [35] F ailure Insensiti v e Routing (FIR) FIR impro v es the a v ailability and stability by suppress- ing the adv ertisement of f ailure paths and tra v erse traf fic along the loop free path. Y ang Richard Y ang et al. [36] Rational Route Selection Algorithms Analysis of set of guidelines for interdomain traf fic engi- neering has been done. If AS follo ws this guidelines, it pro vides guaranteed stable route solution. Shi v ani desh- pande et al. [38] Statistical P attern Recogni- tion T echniques The method performs well under all kinds of instability . In t his, features has been e xtracted from the BGP update messages for capturing the statistical changes. Geof f Huston et al. [39] P ath Exploration Damping (PED) PED Reduces the BGP update message announcement compare to traditional damping approach. Rajvir Gill et al. [41] FLD-MRAI (Fle xible Load Dispersing MRAI) Impro v es con v er gence time by Dispersing netw ork loads o v er routers. Chi Harold Liu. [42] Generic Admission Control (GA C) By controlling the admission for pack et at ingress node the algorithm impro v es the QoS. 2006;1-6. [6] Chin-Fu K uo, Ai-Chun P ang, Sheng-K un Chan, Dynamic Routing with Security Considerations, IEEE T ransa c- tions on P arallel and Distrib uted Systems , January 2009; 20(1): 48-58. [7] Maen Saleh and Liang Dong, Real-T ime Scheduling with Security Enhancement for P ack et Switched Netw orks, IEEE T ransactions on Netw ork and Service Management , September 2013;10(3): 271-285. [8] Sa v erio Mascolo, Smith’ s Principle for Congestion Control in High-Speed Data Netw orks, IEEE T ransactions on Automatic Control , February 2000; 45(2): 558-564. [9] Ian F . Ak yildiz, Giacomo Morabito, and Ser gio P alazzo, TCP-Peach: A Ne w Congestion Control Scheme for Satellite IP Netw orks, IEEE/A CM T ransactions on Netw orking , June 2001; 9(3): 307-321. [10] Atilla Eryilmaz, and R. Srikant, Joint Congestion Control, Routing, and MA C for Stability and F airness in W ireless Netw orks, IEEE Journal on Selected Areas in Communications , August 2006; 24(8): 1514-1524. [11] Jianhua He, Hsiao-Hw a Chen, Thomas M. Chen, and W enqing Cheng, Adapti v e Congestion Control for DSRC V ehicle Netw orks, IEEE Communications Letters , February 2010; 14(2): 127-129. [12] Haitao W u, Zhenqian Feng, Chuanxiong Guo, Y ongguang Zhang, ICTCP: Incast Congestion Control for TCP in Data-Center Netw orks, IEEE/A CM T ransactions on Netw orking , April 2013; 21(2): 345-358. [13] Y an Zhang, Nirw an Ansari, On Architecture Design, Congestion Notication, TCP Incast and Po wer Consumption in Data Centers, IEEE Communications Surv e ys & T utorials , First Quarter 2013; 15(1): 39-64. [14] Shikhar Shukla, Shing au Chan, Adrian S.-W . T am, Abhishek Gupta, Y ang Xu, and H. Jonathan Chao, TCP PLA T O: P ack et Labelling to Alle viate T ime-Out, IEEE Journal on Selected Areas in Communications , January 2014; 32(1): 65-76. [15] Ferhat Dikbiyik, Massimo T ornatore, and Bisw anath Mukherjee, Exploit ing Excess Capacity for Survi v able T raf- fic Grooming in Optical Backbone Netw orks, Journal of Optical Communication Netw ork , FEBR U AR Y 2014; 6(2): 127-137. [16] http://www .peplink.com/technology/load/balancing - algorithms/. [17] Xiaohua Jia, Xiao-Dong Hu, Lu Ruan,and Jianhua Sun, Multicast Routing, Load Balancing, and W a v elength Assignment on T ree of Rings, IEEE Communications Letters , February 2002; 6(2): 79-81. [18] Kartik Gopalan, Tzi-ck er Chiueh, Y o w-JianLin, Load Balancing Routing with Bandwidth-Delay Guarantees, QoS in IP and W ireless Netw ork, IEEE Communication Mag azine , June 2004; 108-113. [19] Lu Ruan, Haibo Luo, and Chang Liu, A Dynamic Routing Algorithm W ith Load Balancing Heuristics for Issues in Routing Mec hanism for P ac k et F orwar ding : A Surve y (R. N. De vikar) Evaluation Warning : The document was created with Spire.PDF for Python.
430 ISSN: 2088-8708 Restorable Connections in WDM Netw orks, IEEE Journal on Selected Areas in Communications , No v ember 2004; 22(9): 1823-1829. [20] A. K. Mishra and A. Sahoo, S-OSPF: A T raf fic Engineering Solution for OSPF Based on Best Ef fort Netw orks, In Proc. IEEE Globecom 2007; 1845-1849. [21] Jiayue He and Jennifer Re xford, T o w ard Internet-W ide Multipath Routing, IEEE Netw ork , March/April 2008; 16-21. [22] Marija Anti, Aleksandra Smiljani, Routing with Load Balancing: Increasing the Guaranteed Node T rafcs, IEEE Communications Letters , June 2009; 13(6): 450-452. [23] Sangsu Jung, Malaz Ksera wi, Dujeong Lee, and June-K oo K e vin Rhee, Distrib uted Potential Field Based Routing and Autonomous Load Balancing for W ireless Mesh Netw orks, IEEE Communications Letters , June 2009; 13(6): 429-431. [24] Marija Anti, and Aleks andra Smiljani, Cost Reduction of Reliable Netw orks Using Load Balanced Routing, IEEE Communications Letters , March 2010; 14(3): 263-265. [25] Eiji Oki and A yak o Iw aki, Load-Balanced IP Routing Scheme Based on Shorte st P aths in Hose Model, IEEE T ransactions on Communications , July 2010; 58(7): 2088-2096. [26] Brice Augustin, T imur Friedman, and Renata T eix eira, Measuring Multipath Routing in the Internet, IEEE/A CM T ransactions on Netw orking , June 2011; 19(3): 830-840. [27] Geor ge Athanasiou, K ostas Tsagkaris, P anagiotis Vlacheas, Dimitrios Karv ounas, and P anagiotis Demestichas, Multi-Objecti v e T rafc Engineering for Future Netw orks, IEEE Communications Letters , January 2012; 16(1): 101-103. [28] Fung Po Tso and Dimitrios P . Pezaros, Impro ving Data Center Netw ork Utilization Using Near -Optimal T raf fic Engineering, IEEE T ransaction on P arallel and Distrib uted System , June 2013; 24(6): 1139-1147. [29] Shuo F ang, Y ang Y u, Chuan Heng F oh, and Khin Mi Mi Aung, A Loss-Free Multipathing Solution for Data Center Netw ork Using Softw are-Defined Netw orking Approach, IEEE T ransaction on Magnetics , June 2013; 49(6): 2723-2729. [30] W idhi Y ah ya, Achmad Basuki, Jehn-Rue y Jiang, International Journal of Electrical and Computer Engineering (IJECE) , April 2015; 5(2): 289-296. [31] Craig Labo vitz, G. Robert Malan, and F arnam Jahanian, Internet Routing Instability , IEEE/A CM T ransactions on Netw orking , October 1998; 6(5): 515-528. [32] Aristotelis Tsirigos and Zygmunt J. Haas, Multipath Routing in the Presence of Frequent T opological Changes, IEEE Communications Mag azine , No v ember 2001; 132-138. [33] P aolo Narv ez, Kai-Y eung Siu, and Hong-Y i Tzeng, Ne w Dynamic Algorithms for Shortest P ath T ree Computa- tion, IEEE/A CM T ransactions on Netw orking , December 2000; 8(6): 734-746. [34] P aolo Narv ez, Kai-Y eung Siu, and Hong-Y i Tzeng, Ne w Dynamic SPT Algorithm Based on a Ball-and-String Model, IEEE/A CM T ransactions on Netw orking , December 2001; 9(6): 706-718. [35] Srihari Nelakuditi, Sanghw an Lee, Y inzhe Y u, Zhi-Li Zhang, and Chen-Nee Chuah, F ast Local Rerouting for Handling T ransient Link F ailures, IEEE/A CM T ransactions on Netw orking , April 2007; 15(2): 359-372. [36] Y ang Richard Y ang, Haiyong Xie, Hao W ang, A vi Silberschatz, Arvind Krishnamurth y , Y anbin Liu, Li Erran Li, On Route Selection for Interdomain T raf fic Engineering, IEEE Netw ork , No v ember/December 2005; 20-27. [37] W ei Sun, Zhuoqing Morle y Mao, Kang G. Shin, Dif ferentiated BGP Update Processing for Impro v ed Routing Con v er gence, IEEE Conference Publication, ICNP Netw ork Protocol , 12-15 No v ember 2006. [38] Shi v ani Deshpande, Marina Thottan, T inKam Ho , and Biplab Sikdar , An Online Mechanism for BGP Instability Detection and Analysis, IEEE T ransactions on Computers , No v ember 2009; 58(11): 1470-1484. [39] Geof f Huston, Mattia Rossi, and Gren ville Armitage, A T echnique for Reducing BGP Update Announcements through P ath Exploration Damping, IEEE Journal on Selected Areas in Communications , October 2010; 28(8); 1271-1286. [40] Mohammad Y anuar Hariya w an, Comparison Anal ysis of Reco v ery Mechanism at MPLS Netw ork, International Journal of Electrical and Computer Engineering(IJECE) , December 2011; 1(2): 151-160 [41] Rajvir Gill, Ra vinder P aul, and Ljiljana T rajk o vic, Ef fect of MRAI T imers and Routing Policies on BGP Con- v er gence T imes, In proc. IPCCC, IEEE 31st International Conference , 1-3 December 2012. [42] Chi Harold Liu, Kin K. Leung, and Athanasios Gk elias, A Generic Admission-Control Methodology for P ack et Netw orks”, IEEE T ransactions on W ireless Communications , February 2014; 13(2): 604-617. IJECE V ol. 6, No. 1, February 2016: 421 430 Evaluation Warning : The document was created with Spire.PDF for Python.