TrieC: A High-Speed IPv6 Lookup with Fast Updates Using Network Processor

Size: px
Start display at page:

Download "TrieC: A High-Speed IPv6 Lookup with Fast Updates Using Network Processor"

Transcription

1 TrieC: High-Speed IPv6 Lookup with Fast Updates Using Network Processor Xianghui Hu 1, Bei Hua 1, and Xinan Tang 2 1 Department of Computer Science and Technology, University of Science and Technology of China, Hefei, P.R. China xhhu@mail.ustc.edu.cn, bhua@ustc.edu.cn 2 Intel Compiler Lab, US xinan.tang@intel.com bstract. ddress lookup is one of the main bottlenecks in Internet backbone routers, as it requires the router to perform a longest-prefix-match when searching the routing table for a next hop. Ever-increasing Internet bandwidth, continuously growing prefix table size and inevitable migration to IPv6 address architecture further exacerbate this situation. In recent years, a variety of highspeed address lookup algorithms have been proposed, however most of them are inappropriate to IPv6 lookup. This paper proposes a high-speed IPv6 lookup algorithm TrieC, which achieves the goals of high-speed address lookup, fast incremental prefix updates, high scalability and reasonable memory requirement by taking great advantage of the network processor architecture. Performance of TrieC is carefully evaluated with several IPv6 routing tables of different sizes and different prefix length distributions on Intel IXP2800 network processor(npu). Simulation shows that TrieC can support IPv6 lookup at OC-192 line rate. Furthermore, if TrieC is pipelined in hardware, it can achieve one IPv6 lookup per memory access. Keywords: Network processor, IPv6 lookup, parallel programming, embedded system design, routing, prefix expansion. 1 Introduction Due to the rapid growth of the Internet bandwidth and continuously increasing size of the routing tables, IP address lookup becomes one of the most challenging tasks in backbone routers. By the inevitable migration to the next generation IPv6 128-bit address space, IPv6 lookup becomes even more demanding. Traditional routers generally use application specific integrated circuits (SIC), FPG, or general-purpose processor (GPP) as a building block. SIC provides guaranteed high-performance with low power, but lack of flexibility makes it unable to keep up with the rapid changes in network protocols. FPG offers certain flexibility but it costs more to build and its power consumption is very high. On the other hand, GPPs can meet the requirements of flexibility, short development period and low cost, but often fail to meet the performance requirements because they are not specially optimized for network processing. For example, the efficiency of the GPP cache system relies on data s temporal locality, which is not a common case in L.T. Yang et al. (Eds.): ICESS 2005, LNCS 3820, pp , Springer-Verlag Berlin Heidelberg 2005

2 118 X. Hu, B. Hua, and X. Tang today s high-speed and aggregated networks. s a consequence, network processor unit(npu) emerges as a promising candidate for networking building block. It retains both the high performance of SIC and flexibility advantage of GPP through parallel and programmable architecture. t present, many companies including Intel, Freescale, grere, MCC and EZchip have developed programmable NPUs. Many system companies including Cisco, lcatel, HUWEI, and ZTE use NPUs to build switches and routers. This paper presents an efficient IPv6 address lookup scheme called TrieC, which achieves O(1) search time with fast incremental updates and reasonable memory requirement by taking great advantage of the characteristics of NPU, especially of Intel IXP network processor. lthough our experiment was done on Intel IXP2800, the same performance can be achieved on other similar NPUs. The main contributions of the paper are as follows: new IPv6 address lookup algorithm (TrieC) is proposed with the features of high speed, fast prefix incremental updates, high scalability and reasonable memory requirement. modified compact prefix expansion (MCPE) technique is designed to use less memory for address search and prefix incremental update than traditional prefix expansion. n architecture awareness algorithm implementation aiming at IXP2800 NPU is elaborated, which: 1) takes advantage of the special instruction set of IXP 2800, especially the bit counting and CRC instruction to make the search of the compressed tables fast; 2) distributes IPv6 routing table in four SRM channels to support simultaneously data accesses; and 3) partitions the tasks appropriately on three IXP2800 Microengines(MEs) to achieve IPv6 lookup at OC-192 line rate. The rest of the paper is organized as follows. Section 2 describes issues of existing approaches in IP address lookup, especially in the IPv6 circumstances. Section 3 explains the design and mechanism of TrieC. Section 4 discusses how to implement incremental prefix updates efficiently. Section 5 introduces the optimized implementation on IXP2800. Section 6 shows simulation results and performance analysis. Finally, section 7 concludes. 2 Related Work The most popular data structure for longest prefix match is trie[6-8]. In order to reduce memory accesses in trie, various kinds of techniques such as prefix expansion and multibit trie[12] have been proposed. Multi-bit trie expands a set of arbitrary length prefixes to a predefined set of prefixes by prefix expansion. Its search time is linear with the multi-bit tree levels and its update time depends on both prefix length and maximum node size. However, its worst-case memory requirement is O(2 k *N*W/k), where k, N and W are search stride, number of prefixes and maximum prefix length respectively. Basic-24-8-DIR[13] is a hardware implementation using prefix expansion for IPv4 lookup with maximum two memory accesses, but it needs more than 32Mbytes memory and even more memory or dual memory bank for routing updates.

3 TrieC: High-Speed IPv6 Lookup with Fast Updates Using Network Processor 119 Waldvogel et al.[9] use binary search on hash table organized by prefix length. The scheme requires log 2 W memory accesses, where W is the maximum prefix length, in the worst case. However, it requires very long preprocessing time to compute markers noting the existence of longer prefixes, hence the update time is O(N*log 2 W) and the whole routing table must be reconstructed. Multiway range tree [11] reduces search time and update time to O(k*log k N) through modifying binary search by prefix length, it also analyzes the feasibility for IPv6 address lookup. However, its memory requirement is O(k*N*log k N). Lapson et al.[10] introduce multicolumn search for IPv6 addresses that avoided the multiplicative factor of W/M inherent in basic binary search by doing binary search in columns of M bits, and moving between columns using pre-computed information. However, in the worst case, it needs approximate 15 memory accesses for IPv6 address lookup because of O(log k 2N+W/M) search time. dditionally, TCM-based, CPU caching[5], reconfigurable fast IP lookup engine[14], binary decision diagrams[20] etc. are all hardware-based IP lookup schemes. Their advantages are high lookup speed, whereas their disadvantages like specific hardware support, high power consumption, complicated prefix update and high cost limit their application to a certain degree. Obviously, the main problem of existing lookup schemes is that they cannot combine high-speed lookup, fast updates and acceptable memory storage for IPv6 address lookup at the same time. This paper presents our solution of an IPv6 lookup scheme-triec based on the modified compact prefix expansion (MCPE) technique and fixed-level multibit trie structure. High performance search is achieved through fixing the levels of TrieC tree, and fast incremental update is achieved by storing the unexpanded prefix length in MCPE nodes. Moreover, memory requirement is reduced to a reasonable capacity due to compressed MCPE technique. 3 lgorithm Design and Mechanism 3.1 Basic Idea In trie structure, prefix information is stored along the path from the root to the leaf node of the tree. To reduce the path length and thus memory access times, prefix expansion technique is applied to increase the routing table size in a fixed stride so that the resulted expanded table could be visited in the same stride index. The proposed scheme is based on the following observations: 1. 2 n-m redundant next-hop information must be stored if an m-bit prefix is expanded to 2 n n-bit prefixes, where n is equal to or greater than m, using prefix expansion. 2. Statistics of existing IPv6 routing tables and the IPv6 addresses allocation policies indicate that the percentage of the prefixes whose lengths are equal to or greater than 48-bit is approximate only 5%. 3. Only aggregatable global unicast addresses, whose format prefix (FP) field is always set to 001, need to be searched in the allocated IPv6 address space. dditionally, the lower 64 bits of IPv6 address are allocated to interface ID, so the core router can ignore them.[1-3]

4 120 X. Hu, B. Hua, and X. Tang Therefore, the basic idea of TrieC is to ignore the highest three bits, build a four-level compressed multibit trie tree using the stride for the prefixes whose lengths are longer than 3bits and shorter than 49bits, then use hash to search the other prefixes whose lengths are longer than 48bits and shorter than 65bits. 3.2 Modified Compact Prefix Expansion The modified compact prefix expansion (MCPE) technique is motivated by the fact that there is a lot of redundant next-hop information in traditional prefix expansion. For example, if the IPv6 prefix (2002:4*::/18,) and (2002:5*::/20,B) are expanded to 24-bit prefixes using the traditional prefix expansion, 64(= ) new prefixes are formed as shown in Fig.1(a). Obviously, the same next-hop index repeats many times. repeats 48 times totally, and B repeats 16 times. 24-bit Index 2002:40*::/ :41*::/ :4F*::/ :50*::/ :51*::/ :5F*::/ :60*::/ :61*::/ :6F*::/ :70*::/ :71*::/ :7F*::/24 Next-Hop B B B (a) Traditional prefix expansion 18-bit Tindex 6-bit Bindex :4*::/ NHI B nul (b) Modified compact prefix expansion Fig. 1. Traditional prefix expansion vs. modified compact prefix expansion (MCPE) The main idea of MCPE is to store consecutively identical next-hop index only once in a next-hop index array (NHI). The NHI in Fig.1(b) has three entries (, B, ); the highest 18 bits of 24-bit prefix are used as the index to search a Tindex table and the next 6 bits are used as another index to search a bit-vector Bittlas in a Bindex table. The 64-bit Bittlas is organized as follows: if the next-hop information denoted by bit I is the same as that denoted by bit I-1, bit I is set to 0; otherwise, bit I is set to 1, indicating that different next-hop information must be added into NHI. In Fig. 2 (b), bit 0 is 1 since it starts a new NHI entry; bit 16 is 1 since its NHI is B that is different from previous entry ; bit 32 is 1 since its NHI is again which is different from previous entry B. Now assume we want to search the next-hop information relating to IPv6 address 2002:6*::/24. Firstly, we use the highest 18 bits as Tindex to find out the MCPE entry 2002:4*::/18, then we use the next 6 bits as Bindex to find the bit offset in Bittlas, which is 42. Since total three bits are set in Bittlas from offset 0 to offset 42, the third element in NHI is the lookup result. The TrieC table in Fig.1 is called TrieC18/6. Similarly, TrieCm/n is designed to represent 2 (m+n) uncompressed (m+n)-bit prefixes. Using MCPE technique, the TrieC tree eliminates redundant information and preserves the high-speed index access characteristic of traditional prefix expansion technique.

5 TrieC: High-Speed IPv6 Lookup with Fast Updates Using Network Processor Data Structure Our stride series for TrieC algorithm is The data structures include three types of tables: TrieC15/6 table (ignored highest three bits 001), TrieC4/4 table, and Hash-16 table. TrieC15/6 table is the first-level table that stores all the prefixes whose lengths fall into [1:24]-bit. TrieC4/4 tables are from the second to the fourth level of TrieC trees, whose prefix lengths belong to [25:32]-bit, [33:40]-bit, and [41:48]-bit respectively. Hash16 table stores all the prefixes whose lengths belong to [49:64]-bit. 1bit 9bits 6bits 0 Next-Hop ID Prefix Length 1 Index into Next Level TrieC (a) Next-Hop Index Bittlas Next-Hop Index 1 Next-Hop Index 2 Next-Hop Index 3 Next-Hop Index 4 (b) Basic TrieC15/6 entry Bittlas Next-Hop Index 1 Next-Hop Index 2 Next-Hop Index 3 (d) Basic TrieC4/4 entry NHI Bittlas NHI 2. Index into ExtraNHI table NHI TotalPosition-1 Reserved NHI TotalPosition (c) TrieC15/6 entry with ExtraNHI NHI NHI Bittlas Reserved. Index into ExtraNHI table NHI TotalPosition-1 NHI TotalPosition (e) TrieC4/4 entry with ExtraNHI Fig. 2. Data structures of TrieC scheme The next-hop index (NHI) structure, which stores the lookup result including the next-hop IP address and the output interface, is shown in Fig. 2(a). The original prefix length is stored in NHI due to the requirement of incremental prefix updates. Each NHI entry is 2 bytes, with the most significant bit setting to 0 indicating that the remaining bits consist of a next-hop ID in NHI[14:6], and an unexpanded prefix length in NHI[5:0], while a 1 in this bit indicating that the remaining 15 bits contain a pointer to the next level TrieC node. The TrieC15/6 table contains 2 15 entries, which is called TrieC15/6_entry, and has two types of structures: Basic and ExtraNHI. The basic structure supports up to four NHIs and ExtraNHI supports more NHIs, in which: 1. TrieC15/6_entry [127:64]: stores a 64-bit vector Bittlas. The least significant bit is always set to one because each IP address absolutely matches the default route. TotalEntropy that is the total number of bits set in the bit vector represents the size of array NHI or ExtraNHI. For a bit position P, the number of bits set in Bittlas[P:0] named PositionEntropy[P] gives the NHI index in array NHI or ExtraNHI. For each P, the equation PositionEntropy[P]<=TotalEntropy always holds. 2. TrieC15/6_entry[63:0]: stores up to 4 NHIs or a pointer to ExtraNHI array. If TotalEntropy of Bittlas field is no greater than 4, TrieC15/6_entry[63:0] stores NHI1, NHI2, NHI3 and NHI4 orderly as shown in Fig. 2(b). Otherwise, TrieC15/6_entry[63:32] stores a 32-bit pointer that points to ExtraNHI array as shown in Fig. 2(c). Similarly, each TrieC4/4 table has 2 4 entries and each entry is 8 bytes. The structures of basic and ExtraNHI TrieC4/4 entry are shown in Fig. 2 (d) and (e). The

6 122 X. Hu, B. Hua, and X. Tang unique difference among all TrieC4/4 tables is that if the flag bit of NHI in the fourth level of TrieC tree is set to one, the Hash16 table must be searched. The Hash16 table uses cyclic redundancy check (CRC) as a hash function that is known as a semiperfect hash function. The structure of a Hash16 entry is (prefix, next-hopid) pair. 3.4 Lookup Mechanism Fig. 3 gives a routing table search algorithm based on compressed TrieC tables. We will use an example to show how these tables are searched. IPv6_Lookup_TrieC (IN DstIP, OUT Next-HopID) { 1. Current_Block = TrieC15_6; 2. Tindex = DstIP [124:110]; 3. Bit_Vec = GetBitVec (Current_Block, Tindex); 4. Bindex = DstIP [109:104]; 5. NHI = GetNHI(Bit_vec, Bindex); 6. if (NHI.flag = 0) return NHI.Next-HopID; 7. else 8. {// search TrieC4/4 tables, base[i] is base of (i+1) th -level TrieC tree 9. Current_Block = TrieC4/4 at Base[0]+NHI[14:0]; 10. for (i=1;i<=3;i++) 11. { 12. Tindex = DstIP[103-8*(i-1):100-8*(i-1)]; 13. Bit_vec = GetBitVec (Current_Block, Tindex); 14. Bindex= DstIP[99-8*(i-1):96-8*(i-1)]; 15. NHI = GetNHI (Bit_Vec, Bindex); 16. If (NHI.flag = 0) return NHI.Next-HopID; 17. else 18. { 19. if (i!=3) Current_Block=TrieC4/4 at Base[i]+NHI [14:0]<<4; 20. else break; //search longer prefix in Hash } 22. } 23. if (Hash (DstIP [79:64])) return Next-HopID; 24. else return Default-Next-HopID; 25. } }// IPv6_Lookup_TrieC Fig. 3. Pseudo code to search TrieC multi-level tree for IPv6 address ssume that the following routes are already in the TrieC table: (2002:4C60::/18,), (2002:4C6F::/28,B). The first route requires an entry in TrieC15/6 that corresponds to the 24-bit prefixes from 2002:40*::/24 to 2002:7F*::/24. The second route further needs a second level TrieC4/4 to be used because its length is 28-bits. Suppose we are looking up the destination IPv6 ddress 2002:4C6::200C, Fig. 4 shows the detailed lookup process. Firstly, DstIP[124:110] that is is used as the Tindex to find the entry 2002:4*::/18 in TrieC15/6; then DstIP[109:104] that is is used as the Bindex to find the bit offset that is 12 in Bittlas; two bits set from offset 0 to offset 12 makes PositionEntropy=2, thus the second entry in NHI is located; a 1 in NHI[15] indicates that NHI[14:0] contains a pointer to the

7 TrieC: High-Speed IPv6 Lookup with Fast Updates Using Network Processor 123 next level TrieC4/4, so NHI[14:0]<<4+DstIP[103:100] is used as the Tindex to the next level TrieC4/4, and then DstIP[99:96] is used as the Bindex to find out the PositionEntropy that is 1, which then locates the desired next-hop ID, which is B. 24 bits 8 bits 8 bits 8 bits 16 bits Interface ID x200C Bindex 0110 Bindex PositionEntropy=2 TrieC4/4 PositionEntropy=1 Tindex 0 B Tindex Index Next-hop ID=B TrieC15/6 NHI[14:0]<<4+DstIP[103:100] Example Routing Table: Destination IPv6ddress 2002:4C6::200C 2002:4C60::/ :4C6F::/28 B 4 Routing Updates Fig. 4. IPv6 address lookup example of the TrieC scheme New routing information must be exchanged among routers as network topology changes. The frequency of routing updates could be as high as a few hundred times per second. Since the routing table cannot be accessed during update, routing updates must be executed fast and efficiently. The operation of adding a new prefix, NewRoute(new_prefix/new_length, new_next-hop), can be classified into two categories according to the value of new_length: Bittlas affected and Bittlas unaffected. The range of the prefix length corresponding to the former is called TableRange[start_length, end_length] and the range corresponding to the latter is called BittlasRange[start_length, end_length]. For instance, the TableRange and BittlasRange of TrieC15/6 table are [4,18] and [19,24] respectively. Consider the updates in a TableRange. Because a NewEntry matches 2 (TableRange.end_length-new_length) TrieC entries and their whole Bittlas fields, we only need update the NHIs of matched entries. If an old NHI satisfies: flag=0, PrefixLength<=NewEntry.new_length and Next-HopID NewEntry.new_next-hop, it is replaced by the NewEntry. Otherwise, we need search and update the next level of TrieC tree. In the case of BittlasRange, NewEntry exactly matches one TrieC entry and 2 (BittlasRange.end_length-new_length) bits in the Bittlas field of the matched entry. For each matched bit, we only need update the bit and its corresponding NHI if PrefixLength<=NewEntry.new_length and Next-HopID NewEntry.new_next-hop. Otherwise, we need search and update the next level of TrieC tree. The process of adding new prefixes (2002:E7B::/17,C) and (2002:279::/23,D) into the example routing table is shown as Fig. 5. The prefix (2002:E7B::/17,C) matches two TrieC15/6 entries with the Tindex and Only the first NHI() of the former TrieC15/6 entry is replaced by NHI(0,C,17) because the NHI() matches the update condition. In Fig. 5(b), the prefix

8 124 X. Hu, B. Hua, and X. Tang C Routing Table 2002:4C60::/ :4C6F::/28 B 2002:E7B::/17 C 2002:279::/23 D 0B TrieC15/6 (a) TrieC tree after (2002:E7B::/17,C) added F TrieC4/ C 17 0 D C B TrieC15/6 (b) TrieC tree after (2002:279::/23,D) added F TrieC4/4 Fig. 5. Routing updates of the example IPv6 routing table (2002:279::/23,D) matches only one TrieC15/16 entry, whose Tindex is , and two bits in the Bittlas field of this entry. Because the NHI corresponding to these two bits matches the update condition, i.e., NHI.Prefix- Length=17<new_length=23 and NHI.Next-HopID=C new_next-hop=d, both of these two bits are set to one and the corresponding NHIs are updated. 5 Optimization on IXP2800 Network Processor Intel IXP2800 NPU is a programmable network processor that comprises a single XScale processor, sixteen Microengines(MEs), four SRM controllers, three RDRM controllers and high-speed bus interfaces. Each ME has eight hardwareassisted threads of execution. ll threads in a particular ME execute the same code stored on that ME. Complete descriptions of the hardware architecture and software framework are available from the IXP2800 manuals[19]. We implemented TrieC algorithm in MicroengineC language and simulated it on Intel Developer DevWorkbench 4.1[19], which offers a cycle-accurate simulator of IXP2800 network processor. Through analysis, we identified the following operations as time-consuming ones: calculation of TotalEntropy and PositionEntropy, SRM memory accesses, and CRC hash operation. The efficiency of those operations affects the performance of TrieC greatly. We will show how to optimize these operations by taking advantage of the characteristics of IXP2800 network processor. Firstly, the main workload of entropy calculation in TrieC is to count the number of bit set in bit-vector Bittlas. The naive implementation of checking each bit and counting the number of bit set seen so far it is very slow and needs a few hundred shift, LU and branch instructions. Intel IXP2800 network processor has a built-in instruction POP_COUNT that can calculate the number of bit set in a 32-bit register in three clock cycles. Such tremendous reduction of the number of instructions lays a solid foundation for TrieC to achieve line rate.

9 TrieC: High-Speed IPv6 Lookup with Fast Updates Using Network Processor 125 On IXP2800 NPU, each SRM access takes more than 100 cycles, hiding memory latency becomes another important step towards high-performance. We hid the memory-access latencies by means of: Parallelized access: IXP2800 NPU contains four independent SRM controllers and each channel supports up to 64Mbytes SRM. We partitioned the TrieC tables, and then distributed them properly into four SRM channels. TrieC15/6 table and hash16 table were stored in SRM channel 0, because their sizes were smaller than that of each TrieC4/4. The second, third and fourth level TrieC4/4 tables were stored in SRM channel 1, 2 and 3 respectively, so that they could be accessed in parallel. Vectoring access: On IXP2800 NPU, adjacent SRM locations can be fetched in one instruction to the array of SRM transfer registers, which allows for a significant reduction in the number of SRM access. s the maximal length of each SRM access instruction can be up to 64-bytes, we carefully designed the data structures of Tries15/6 and TriesC4/4 such that the sizes of TrieC15/6 entry and TrieC4/4 entry were less than 64-bytes and thus each table entry could be fetched in one SRM access. Finally, each ME has a CRC Unit, which operates in parallel with the execution of data path and supports one time CRC operation within the period of each two consecutive instructions. By using the CRC unit to perform hash calculation, TrieC speeded up the search of hash16 table. 6 Simulation and Performance nalysis Since IPv6 is not yet widely deployed, existing IPv6 tables, which normally have less 1000 prefixes[16][17], are small and unlikely to reflect future IPv6 network growth. Currently, randomly generated tables are often used for IPv6 research and development. To reflect the IPv6 address distribution as objectively as possible, we used three different ways to generate nine IPv6 routing tables whose prefix length distributions are shown in Tab. 1. Group was generated from the CERNET[17], 6Bone, 6Net and Telstra BGP IPv6 routing tables[16], reflecting the existing IPv6 prefix length distribution. Group B was generated from the non-random generator IPv6 table proposed by M Wang et. al[18], reflecting the ideal IPv6 routing tables. Group C was calculated as the arithmetical average of and B, reflecting the future IPv6 tables. Table 1. Prefix length and entries used in simulation Prefix Number N= N= N= Length B C B C B C

10 126 X. Hu, B. Hua, and X. Tang 40 Group Group B Group C DIR-24-8-BSIC(IPv4) Memory requirement(mb) K 300K 400K Number of prefix Fig. 6. Memory requirements of nine IPv6 routing tables vs. of DIR-24-8-BSIC(IPv4) For each group, we generated three different sizes of tables with 200K, 300K and 400K entries. ll the prefix values are generated randomly. Memory requirements of the nine IPv6 tables are shown in Fig. 6. Obviously, memory requirements increase along with table sizes. Note that the memory requirement of table B-400K is approximate 35Mbytes,which is slightly higher than 33Mbyes of DIR-24-8-BSIC for IPv4, but is significantly lower than the estimated memory requirement of multibit-trie, which is approximately more than 820Mbytes at 8-bit strides for 400K entries. With such high a compression rate, the entire routing tables can be stored in SRM verage memory accesses K B-200K C-200K -300K B-300K C-300K -400K B-400K C-400K Simulation routing table Fig. 7. verage memory accesses of nine IPv6 routing tables In the worst case, TrieC needs eight memory accesses and one hash operation. It is clear from Fig. 7 that there is no any relation between average memory accesses and size of IPv6 routing table; the average memory accesses depend only on the prefix length distribution of the table. For example, the average memory accesses of the three tables belonging to group-b are all close to four, because the percentages of the prefixes whose lengths between bits in these tables are all higher than 70%. On average, the number of memory access of these nine IPv6 tables is far less than eight, which indicates that the percentage of the ExtraNHI nodes is extremely low. Simulation results demonstrated it, and found that the percentage was only about 3.6%.

11 TrieC: High-Speed IPv6 Lookup with Fast Updates Using Network Processor 127 Table 2. Minimal threads number required to meet OC-192 line rate on IXP2800 NPU Routing table Minimal Threads Lookup rate (Mlps.) number Single-thread Multi-threads Group Group B Group C Worst case ssume that the minimal packet size is 60-bytes, then approximate 20.83Mlps lookup rate is required to support OC-192 line rate. Tab. 2 gives the minimal number of threads required to reach the line rate speed for each group, in which on average 9, 17 and 11 threads are needed for group, B and C respectively. Considering that there are 16 MEs on IXP2800, and TrieC just consumed less than three MEs of entire ME budget; therefore, there is still enough room for other networking application such as classification and traffic management to meet their line rate requirements. Finally, we compare the performance of TrieC with some existing outstanding schemes in Tab. 3. Table 3. Comparison of search, update, space complexities and scalability for IPv6 Scheme Search time Update time Memory requirement IPv6 Patricia trie O(W) O(W) O(N*W) N Multibit trie O(W/k) O(W/k+2 k ) O(2 k *N*W/k) N Binary search O(log 2W) O(N*log 2W) O(N*log 2W) Y Multiway search O(log K2N) O(N) O(N) Y DIR-24-8-BSIC O(1) Dual bank memory 33MB@IPv4 N TrieC O(1) O(W/k) O(N*W/k) Y 7 Conclusion This paper proposes an IXP2800-based high performance IPv6 lookup algorithm (TrieC) that features high-speed address lookup, fast routing table update, high scalability, and reasonable memory requirement. modified compact prefix expansion (MCPE) technique that supports faster address search and prefix incremental update with less memory requirement than traditional prefix expansion is designed to build a four-level TrieC tree for IPv6 address lookup. Techniques such as distributed routing table allocation in four SRM channels, functional pipelining, Bittlas field calculation speedup, parallelized SRM accesses, consolidating adjacent SRM accesses and hardware CRC hash unit are used to optimize the proposed scheme on IXP2800 network processor. These optimizaiton techniques can also be applied to other similar NP architectures. Performance of TrieC is evaluated with nine IPv6 routing tables of different sizes and different prefix length distributions on Intel IXP2800 network processor. Simulation shows that TrieC implemented on IXP2800 can support IPv6 lookup at OC-192 line rate. Furthermore this algorithm can be easily implemented in a pipelined architecture (SIC) and achieve one IPv6 lookup per SRM access.

12 128 X. Hu, B. Hua, and X. Tang References 1. R. Hinden, S. Deering, RFC2373 IP Version 6 ddressing rchitecture 2. R. Hinden, S. Deering, RFC2374 IPv6 ggregatable global unicast address format, 3. S. Deering, R. Hinden, RFC2460 Internet Protocol, Version 6IPv6Specification 4. Internet Performance Measurement and nalysis Project, 5. M.. Ruiz-Sanchez, E.W. Biersack, and W. Dabbous, Survey and taxonomy of IP address lookup algorithms, IEEE Network, vol. 15, pp.8 23, Mar.-pr Morrison, PTRICI-Practical lgorithm to Retrieve Information Coded in lphanumeric, J. CM, vol. 15, no. 4, pp , Oct W. Doeringer, G. Karjoth and M. Nassehi, Routing on Longest Matching Prefixes, IEEE Trans. on Networking, vol. 4, no. 1, pp.86-97, Feb N. Yazdani and P. S. Min, Fast and scalable schemes for the IP address lookup problem, in Proc. IEEE Conf. High Performance Switching and Routing, 2000, pp M. Waldvogel, G. Varghese, J. Turner, and B. Plattner, Scalable high speed IP routing lookups, Proc. CM SIGCOMM 97, vol. 27, no. 4, pp , Oct B. Lampson, V. Srinivasan, and G. Varghese, IP lookups using multiway and multicolumn search, in Proc. IEEE INFOCOM 98, San Francisco, C, 1998, pp Subhash Suri; Varghese, G.; Warkhede, P.R., Multiway Range Trees: Scalable IP Lookup with Fast updates, Global Telecommunications Conference, 2001.GLOBECOM'01. IEEE Volume 3, Nov pp vol V. Srinivasan, G. Varghese, Fast address lookups using controlled prefix expansion, Proc. CM Sigmetrics 98, June 1998 pp P. Gupta, S. Lin, and N. McKeown, Routing lookups in hardware at memory access speeds, in Proc. IEEE INFOCOM 98, vol. 3, San Francisco, US, 1998, pp E. Taylor, J. W. Lockwood, T. S. Sproull, J. S. Turner, and D. B. Parlour, Scalable IP lookup for programmable routers, in Proc. IEEE INFOCOM 2002, vol. 2, New York, 2002, pp Raj Jain, Comparison of Hashing Schemes for ddress Lookup in Computer Networks, Communications, IEEE Transactions on Volume 40, Issue 10, Oct pp M Wang, S Deering, T Hain, L Dunn Non-random Generator for IPv6 Tables High Performance Interconnects, Proceedings 12th nnual IEEE Symposium on ug pp Sangireddy, R.; Somani,.K.; High-speed IP routing with binary decision diagrams based hardware address lookup engine, Selected reas in Communications, IEEE Journal on Volume 21, Issue 4, May 2003 pp

IP Address Lookup Using A Dynamic Hash Function

IP Address Lookup Using A Dynamic Hash Function IP Address Lookup Using A Dynamic Hash Function Xiaojun Nie David J. Wilson Jerome Cornet Gerard Damm Yiqiang Zhao Carleton University Alcatel Alcatel Alcatel Carleton University xnie@math.carleton.ca

More information

Scalable Prefix Matching for Internet Packet Forwarding

Scalable Prefix Matching for Internet Packet Forwarding Scalable Prefix Matching for Internet Packet Forwarding Marcel Waldvogel Computer Engineering and Networks Laboratory Institut für Technische Informatik und Kommunikationsnetze Background Internet growth

More information

IP address lookup for Internet routers using cache routing table

IP address lookup for Internet routers using cache routing table ISSN (Print): 1694 0814 35 IP address lookup for Internet routers using cache routing table Houassi Hichem 1 and Bilami Azeddine 2 1 Department of Computer Science, University Center of Khenchela, Algeria

More information

CHAPTER 5 FINITE STATE MACHINE FOR LOOKUP ENGINE

CHAPTER 5 FINITE STATE MACHINE FOR LOOKUP ENGINE CHAPTER 5 71 FINITE STATE MACHINE FOR LOOKUP ENGINE 5.1 INTRODUCTION Finite State Machines (FSMs) are important components of digital systems. Therefore, techniques for area efficiency and fast implementation

More information

Monitoring Large Flows in Network

Monitoring Large Flows in Network Monitoring Large Flows in Network Jing Li, Chengchen Hu, Bin Liu Department of Computer Science and Technology, Tsinghua University Beijing, P. R. China, 100084 { l-j02, hucc03 }@mails.tsinghua.edu.cn,

More information

Data Structures For IP Lookup With Bursty Access Patterns

Data Structures For IP Lookup With Bursty Access Patterns Data Structures For IP Lookup With Bursty Access Patterns Sartaj Sahni & Kun Suk Kim sahni, kskim @cise.ufl.edu Department of Computer and Information Science and Engineering University of Florida, Gainesville,

More information

FAST IP ADDRESS LOOKUP ENGINE FOR SOC INTEGRATION

FAST IP ADDRESS LOOKUP ENGINE FOR SOC INTEGRATION FAST IP ADDRESS LOOKUP ENGINE FOR SOC INTEGRATION Tomas Henriksson Department of Electrical Engineering Linköpings universitet SE-581 83 Linköping tomhe@isy.liu.se Ingrid Verbauwhede UCLA EE Dept 7440B

More information

Efficient IP-Address Lookup with a Shared Forwarding Table for Multiple Virtual Routers

Efficient IP-Address Lookup with a Shared Forwarding Table for Multiple Virtual Routers Efficient IP-Address Lookup with a Shared Forwarding Table for Multiple Virtual Routers ABSTRACT Jing Fu KTH, Royal Institute of Technology Stockholm, Sweden jing@kth.se Virtual routers are a promising

More information

Routing Prefix Caching in Network Processor Design

Routing Prefix Caching in Network Processor Design Routing Prefix Caching in Network Processor Design Huan Liu Department of Electrical Engineering Stanford University, C 9435 huanliu@stanford.edu bstract Cache has been time proven to be a very effective

More information

International Journal of Scientific & Engineering Research, Volume 4, Issue 8, August-2013 1300 ISSN 2229-5518

International Journal of Scientific & Engineering Research, Volume 4, Issue 8, August-2013 1300 ISSN 2229-5518 International Journal of Scientific & Engineering Research, Volume 4, Issue 8, August-2013 1300 Efficient Packet Filtering for Stateful Firewall using the Geometric Efficient Matching Algorithm. Shriya.A.

More information

Scaling 10Gb/s Clustering at Wire-Speed

Scaling 10Gb/s Clustering at Wire-Speed Scaling 10Gb/s Clustering at Wire-Speed InfiniBand offers cost-effective wire-speed scaling with deterministic performance Mellanox Technologies Inc. 2900 Stender Way, Santa Clara, CA 95054 Tel: 408-970-3400

More information

Scalable High-Speed Prefix Matching

Scalable High-Speed Prefix Matching Scalable High-Speed Prefix Matching MARCEL WALDVOGEL IBM Zurich Research Laboratory GEORGE VARGHESE University of California, San Diego JON TURNER Washington University in St. Louis and BERNHARD PLATTNER

More information

AN OVERVIEW OF QUALITY OF SERVICE COMPUTER NETWORK

AN OVERVIEW OF QUALITY OF SERVICE COMPUTER NETWORK Abstract AN OVERVIEW OF QUALITY OF SERVICE COMPUTER NETWORK Mrs. Amandeep Kaur, Assistant Professor, Department of Computer Application, Apeejay Institute of Management, Ramamandi, Jalandhar-144001, Punjab,

More information

The Load Balancing System Design of Service Based on IXP2400 Yi Shijun 1, a, Jing Xiaoping 1,b

The Load Balancing System Design of Service Based on IXP2400 Yi Shijun 1, a, Jing Xiaoping 1,b Advanced Engineering Forum Vol. 1 (2011) pp 42-46 Online: 2011-09-09 (2011) Trans Tech Publications, Switzerland doi:10.4028/www.scientific.net/aef.1.42 The Load Balancing System Design of Service Based

More information

A Fast Pattern-Matching Algorithm for Network Intrusion Detection System

A Fast Pattern-Matching Algorithm for Network Intrusion Detection System A Fast Pattern-Matching Algorithm for Network Intrusion Detection System Jung-Sik Sung 1, Seok-Min Kang 2, Taeck-Geun Kwon 2 1 ETRI, 161 Gajeong-dong, Yuseong-gu, Daejeon, 305-700, Korea jssung@etri.re.kr

More information

Stateful Inspection Firewall Session Table Processing

Stateful Inspection Firewall Session Table Processing International Journal of Information Technology, Vol. 11 No. 2 Xin Li, ZhenZhou Ji, and MingZeng Hu School of Computer Science and Technology Harbin Institute of Technology 92 West Da Zhi St. Harbin, China

More information

Hyper Node Torus: A New Interconnection Network for High Speed Packet Processors

Hyper Node Torus: A New Interconnection Network for High Speed Packet Processors 2011 International Symposium on Computer Networks and Distributed Systems (CNDS), February 23-24, 2011 Hyper Node Torus: A New Interconnection Network for High Speed Packet Processors Atefeh Khosravi,

More information

Binary search tree with SIMD bandwidth optimization using SSE

Binary search tree with SIMD bandwidth optimization using SSE Binary search tree with SIMD bandwidth optimization using SSE Bowen Zhang, Xinwei Li 1.ABSTRACT In-memory tree structured index search is a fundamental database operation. Modern processors provide tremendous

More information

1. How many unique network IDs are there in class A addresses? # of unique hosts?

1. How many unique network IDs are there in class A addresses? # of unique hosts? CS445: IPv4 Addresses In-class activity Names: Part 1: Address Classes Original three classes of IPv4 addresses: A: 0 network (7 bits) host (24 bits) B: 10 network (14 bits) host (16 bits) C: 110 network

More information

IP Lookups Using Multiway and Multicolumn Search

IP Lookups Using Multiway and Multicolumn Search 324 IEEE/ACM TRANSACTIONS NETWORKING, VOL. I, NO. 3, JUNE 1999 IP Lookups Using Multiway and Multicolumn Search Butler Lampson, Venkatachary Srinivasan, and George Varghese, Associate Member, IEEE Abstract-

More information

Quality of Service Routing Network and Performance Evaluation*

Quality of Service Routing Network and Performance Evaluation* Quality of Service Routing Network and Performance Evaluation* Shen Lin, Cui Yong, Xu Ming-wei, and Xu Ke Department of Computer Science, Tsinghua University, Beijing, P.R.China, 100084 {shenlin, cy, xmw,

More information

Design and Implementation of an On-Chip timing based Permutation Network for Multiprocessor system on Chip

Design and Implementation of an On-Chip timing based Permutation Network for Multiprocessor system on Chip Design and Implementation of an On-Chip timing based Permutation Network for Multiprocessor system on Chip Ms Lavanya Thunuguntla 1, Saritha Sapa 2 1 Associate Professor, Department of ECE, HITAM, Telangana

More information

An Active Packet can be classified as

An Active Packet can be classified as Mobile Agents for Active Network Management By Rumeel Kazi and Patricia Morreale Stevens Institute of Technology Contact: rkazi,pat@ati.stevens-tech.edu Abstract-Traditionally, network management systems

More information

Architecture of distributed network processors: specifics of application in information security systems

Architecture of distributed network processors: specifics of application in information security systems Architecture of distributed network processors: specifics of application in information security systems V.Zaborovsky, Politechnical University, Sait-Petersburg, Russia vlad@neva.ru 1. Introduction Modern

More information

A Fast Path Recovery Mechanism for MPLS Networks

A Fast Path Recovery Mechanism for MPLS Networks A Fast Path Recovery Mechanism for MPLS Networks Jenhui Chen, Chung-Ching Chiou, and Shih-Lin Wu Department of Computer Science and Information Engineering Chang Gung University, Taoyuan, Taiwan, R.O.C.

More information

FORWARDING of Internet Protocol (IP) packets is the primary. Scalable IP Lookup for Internet Routers

FORWARDING of Internet Protocol (IP) packets is the primary. Scalable IP Lookup for Internet Routers Scalable IP Lookup for Internet Routers David E. Taylor, Jonathan S. Turner, John W. Lockwood, Todd S. Sproull, David B. Parlour Abstract IP address lookup is a central processing function of Internet

More information

Reconfigurable Architecture Requirements for Co-Designed Virtual Machines

Reconfigurable Architecture Requirements for Co-Designed Virtual Machines Reconfigurable Architecture Requirements for Co-Designed Virtual Machines Kenneth B. Kent University of New Brunswick Faculty of Computer Science Fredericton, New Brunswick, Canada ken@unb.ca Micaela Serra

More information

Configuring a Load-Balancing Scheme

Configuring a Load-Balancing Scheme Configuring a Load-Balancing Scheme Finding Feature Information Configuring a Load-Balancing Scheme Last Updated: August 15, 2011 This module contains information about Cisco Express Forwarding and describes

More information

ECE 578 Term Paper Network Security through IP packet Filtering

ECE 578 Term Paper Network Security through IP packet Filtering ECE 578 Term Paper Network Security through IP packet Filtering Cheedu Venugopal Reddy Dept of Electrical Eng and Comp science Oregon State University Bin Cao Dept of electrical Eng and Comp science Oregon

More information

Influence of Load Balancing on Quality of Real Time Data Transmission*

Influence of Load Balancing on Quality of Real Time Data Transmission* SERBIAN JOURNAL OF ELECTRICAL ENGINEERING Vol. 6, No. 3, December 2009, 515-524 UDK: 004.738.2 Influence of Load Balancing on Quality of Real Time Data Transmission* Nataša Maksić 1,a, Petar Knežević 2,

More information

Table of Contents. Cisco How Does Load Balancing Work?

Table of Contents. Cisco How Does Load Balancing Work? Table of Contents How Does Load Balancing Work?...1 Document ID: 5212...1 Introduction...1 Prerequisites...1 Requirements...1 Components Used...1 Conventions...1 Load Balancing...1 Per Destination and

More information

Exam 1 Review Questions

Exam 1 Review Questions CSE 473 Introduction to Computer Networks Exam 1 Review Questions Jon Turner 10/2013 1. A user in St. Louis, connected to the internet via a 20 Mb/s (b=bits) connection retrieves a 250 KB (B=bytes) web

More information

Advanced Computer Architecture-CS501. Computer Systems Design and Architecture 2.1, 2.2, 3.2

Advanced Computer Architecture-CS501. Computer Systems Design and Architecture 2.1, 2.2, 3.2 Lecture Handout Computer Architecture Lecture No. 2 Reading Material Vincent P. Heuring&Harry F. Jordan Chapter 2,Chapter3 Computer Systems Design and Architecture 2.1, 2.2, 3.2 Summary 1) A taxonomy of

More information

An Algorithm for Performing Routing Lookups in Hardware

An Algorithm for Performing Routing Lookups in Hardware 31 CHAPTER 2 An Algorithm for Performing Routing Lookups in Hardware 1 Introduction This chapter describes a longest prefix matching algorithm to perform fast IPv4 route lookups in hardware. The chapter

More information

Autonomous Fast Rerouting for Software Defined Network

Autonomous Fast Rerouting for Software Defined Network Autonomous ast Rerouting for Software Defined Network 2012.10.29 NTT Network Service System Laboratories, NTT Corporation Shohei Kamamura, Akeo Masuda, Koji Sasayama Page 1 Outline 1. Background and Motivation

More information

Flexible Deterministic Packet Marking: An IP Traceback Scheme Against DDOS Attacks

Flexible Deterministic Packet Marking: An IP Traceback Scheme Against DDOS Attacks Flexible Deterministic Packet Marking: An IP Traceback Scheme Against DDOS Attacks Prashil S. Waghmare PG student, Sinhgad College of Engineering, Vadgaon, Pune University, Maharashtra, India. prashil.waghmare14@gmail.com

More information

Interconnection Networks. Interconnection Networks. Interconnection networks are used everywhere!

Interconnection Networks. Interconnection Networks. Interconnection networks are used everywhere! Interconnection Networks Interconnection Networks Interconnection networks are used everywhere! Supercomputers connecting the processors Routers connecting the ports can consider a router as a parallel

More information

Research on Errors of Utilized Bandwidth Measured by NetFlow

Research on Errors of Utilized Bandwidth Measured by NetFlow Research on s of Utilized Bandwidth Measured by NetFlow Haiting Zhu 1, Xiaoguo Zhang 1,2, Wei Ding 1 1 School of Computer Science and Engineering, Southeast University, Nanjing 211189, China 2 Electronic

More information

A 10-Gbps High-Speed Single-Chip Network Intrusion Detection and Prevention System

A 10-Gbps High-Speed Single-Chip Network Intrusion Detection and Prevention System A 0-Gbps High-Speed Single-Chip Network Intrusion Detection and Prevention System N. Sertac Artan, Rajdip Ghosh, Yanchuan Guo, and H. Jonathan Chao Department of Electrical and Computer Engineering Polytechnic

More information

Internet Protocols Fall 2005. Lectures 7-8 Andreas Terzis

Internet Protocols Fall 2005. Lectures 7-8 Andreas Terzis Internet Protocols Fall 2005 Lectures 7-8 Andreas Terzis Outline Internet Protocol Service Model Fragmentation Addressing Original addressing scheme Subnetting CIDR Forwarding ICMP ARP Address Shortage

More information

Router Architectures

Router Architectures Router Architectures An overview of router architectures. Introduction What is a Packet Switch? Basic Architectural Components Some Example Packet Switches The Evolution of IP Routers 2 1 Router Components

More information

ADVANCED PROCESSOR ARCHITECTURES AND MEMORY ORGANISATION Lesson-12: ARM

ADVANCED PROCESSOR ARCHITECTURES AND MEMORY ORGANISATION Lesson-12: ARM ADVANCED PROCESSOR ARCHITECTURES AND MEMORY ORGANISATION Lesson-12: ARM 1 The ARM architecture processors popular in Mobile phone systems 2 ARM Features ARM has 32-bit architecture but supports 16 bit

More information

A Hybrid Electrical and Optical Networking Topology of Data Center for Big Data Network

A Hybrid Electrical and Optical Networking Topology of Data Center for Big Data Network ASEE 2014 Zone I Conference, April 3-5, 2014, University of Bridgeport, Bridgpeort, CT, USA A Hybrid Electrical and Optical Networking Topology of Data Center for Big Data Network Mohammad Naimur Rahman

More information

Fast Address Lookups Using Controlled Prefix Expansion

Fast Address Lookups Using Controlled Prefix Expansion Fast Address Lookups Using Controlled Prefix Expansion V. SRINIVASAN and G. VARGHESE Washington University in St. Louis Internet (IP) address lookup is a major bottleneck in high-performance routers. IP

More information

Image Compression through DCT and Huffman Coding Technique

Image Compression through DCT and Huffman Coding Technique International Journal of Current Engineering and Technology E-ISSN 2277 4106, P-ISSN 2347 5161 2015 INPRESSCO, All Rights Reserved Available at http://inpressco.com/category/ijcet Research Article Rahul

More information

3D On-chip Data Center Networks Using Circuit Switches and Packet Switches

3D On-chip Data Center Networks Using Circuit Switches and Packet Switches 3D On-chip Data Center Networks Using Circuit Switches and Packet Switches Takahide Ikeda Yuichi Ohsita, and Masayuki Murata Graduate School of Information Science and Technology, Osaka University Osaka,

More information

Accelerating network security services with fast packet classification

Accelerating network security services with fast packet classification Computer Communications 27 (2004) 1637 1646 www.elsevier.com/locate/comcom Accelerating network security services with fast packet classification Shiuhpyng Shieh 1, Fu-Yuan Lee 1, *, Ya-Wen Lin 1 Department

More information

SDN. WHITE PAPER Intel Ethernet Switch FM6000 Series - Software Defined Networking. Recep Ozdag Intel Corporation

SDN. WHITE PAPER Intel Ethernet Switch FM6000 Series - Software Defined Networking. Recep Ozdag Intel Corporation WHITE PAPER Intel Ethernet Switch FM6000 Series - Software Defined Networking Intel Ethernet Switch FM6000 Series - Software Defined Networking Recep Ozdag Intel Corporation Software Defined Networking

More information

InfiniBand Clustering

InfiniBand Clustering White Paper InfiniBand Clustering Delivering Better Price/Performance than Ethernet 1.0 Introduction High performance computing clusters typically utilize Clos networks, more commonly known as Fat Tree

More information

WHITE PAPER. Understanding IP Addressing: Everything You Ever Wanted To Know

WHITE PAPER. Understanding IP Addressing: Everything You Ever Wanted To Know WHITE PAPER Understanding IP Addressing: Everything You Ever Wanted To Know Understanding IP Addressing: Everything You Ever Wanted To Know CONTENTS Internet Scaling Problems 1 Classful IP Addressing 3

More information

Content-Aware Load Balancing using Direct Routing for VOD Streaming Service

Content-Aware Load Balancing using Direct Routing for VOD Streaming Service Content-Aware Load Balancing using Direct Routing for VOD Streaming Service Young-Hwan Woo, Jin-Wook Chung, Seok-soo Kim Dept. of Computer & Information System, Geo-chang Provincial College, Korea School

More information

SIMULATION STUDY OF BLACKHOLE ATTACK IN THE MOBILE AD HOC NETWORKS

SIMULATION STUDY OF BLACKHOLE ATTACK IN THE MOBILE AD HOC NETWORKS Journal of Engineering Science and Technology Vol. 4, No. 2 (2009) 243-250 School of Engineering, Taylor s University College SIMULATION STUDY OF BLACKHOLE ATTACK IN THE MOBILE AD HOC NETWORKS SHEENU SHARMA

More information

Packet Classification for Core Routers: Is there an alternative to CAMs?

Packet Classification for Core Routers: Is there an alternative to CAMs? Packet Classification for Core Routers: Is there an alternative to CAMs? Florin Baboescu, Sumeet Singh, George Varghese Abstract A classifier consists of a set of rules for classifying packets based on

More information

Security-Aware Beacon Based Network Monitoring

Security-Aware Beacon Based Network Monitoring Security-Aware Beacon Based Network Monitoring Masahiro Sasaki, Liang Zhao, Hiroshi Nagamochi Graduate School of Informatics, Kyoto University, Kyoto, Japan Email: {sasaki, liang, nag}@amp.i.kyoto-u.ac.jp

More information

CROSS LAYER BASED MULTIPATH ROUTING FOR LOAD BALANCING

CROSS LAYER BASED MULTIPATH ROUTING FOR LOAD BALANCING CHAPTER 6 CROSS LAYER BASED MULTIPATH ROUTING FOR LOAD BALANCING 6.1 INTRODUCTION The technical challenges in WMNs are load balancing, optimal routing, fairness, network auto-configuration and mobility

More information

FPGA Implementation of IP Packet Segmentation and Reassembly in Internet Router*

FPGA Implementation of IP Packet Segmentation and Reassembly in Internet Router* SERBIAN JOURNAL OF ELECTRICAL ENGINEERING Vol. 6, No. 3, December 2009, 399-407 UDK: 004.738.5.057.4 FPGA Implementation of IP Packet Segmentation and Reassembly in Internet Router* Marko Carević 1,a,

More information

Physical Data Organization

Physical Data Organization Physical Data Organization Database design using logical model of the database - appropriate level for users to focus on - user independence from implementation details Performance - other major factor

More information

Index Terms Domain name, Firewall, Packet, Phishing, URL.

Index Terms Domain name, Firewall, Packet, Phishing, URL. BDD for Implementation of Packet Filter Firewall and Detecting Phishing Websites Naresh Shende Vidyalankar Institute of Technology Prof. S. K. Shinde Lokmanya Tilak College of Engineering Abstract Packet

More information

International Journal of Advanced Research in Computer Science and Software Engineering

International Journal of Advanced Research in Computer Science and Software Engineering Volume 2, Issue 9, September 2012 ISSN: 2277 128X International Journal of Advanced Research in Computer Science and Software Engineering Research Paper Available online at: www.ijarcsse.com An Experimental

More information

System Interconnect Architectures. Goals and Analysis. Network Properties and Routing. Terminology - 2. Terminology - 1

System Interconnect Architectures. Goals and Analysis. Network Properties and Routing. Terminology - 2. Terminology - 1 System Interconnect Architectures CSCI 8150 Advanced Computer Architecture Hwang, Chapter 2 Program and Network Properties 2.4 System Interconnect Architectures Direct networks for static connections Indirect

More information

Network Attached Storage. Jinfeng Yang Oct/19/2015

Network Attached Storage. Jinfeng Yang Oct/19/2015 Network Attached Storage Jinfeng Yang Oct/19/2015 Outline Part A 1. What is the Network Attached Storage (NAS)? 2. What are the applications of NAS? 3. The benefits of NAS. 4. NAS s performance (Reliability

More information

Scalability and Classifications

Scalability and Classifications Scalability and Classifications 1 Types of Parallel Computers MIMD and SIMD classifications shared and distributed memory multicomputers distributed shared memory computers 2 Network Topologies static

More information

IP Router Architectures: An Overview

IP Router Architectures: An Overview IP Router Architectures: An Overview by James Aweya {Email: aweyaj@nortelnetworks.com; Tel: 1-613-763-6491; Fax: 1-613-763-5692} Nortel Networks Ottawa, Canada, K1Y 4H7 Abstract In the emerging environment

More information

On Characterizing BGP Routing Table Growth Tian Bu, Lixin Gao, and Don Towsley University of Massachusetts, Amherst, MA 01003

On Characterizing BGP Routing Table Growth Tian Bu, Lixin Gao, and Don Towsley University of Massachusetts, Amherst, MA 01003 On Characterizing BGP Routing Table Growth Tian Bu, Lixin Gao, and Don Towsley University of Massachusetts, Amherst, MA 0003 Abstract The sizes of the BGP routing tables have increased by an order of magnitude

More information

Understanding the Benefits of IBM SPSS Statistics Server

Understanding the Benefits of IBM SPSS Statistics Server IBM SPSS Statistics Server Understanding the Benefits of IBM SPSS Statistics Server Contents: 1 Introduction 2 Performance 101: Understanding the drivers of better performance 3 Why performance is faster

More information

Scalable Internet Services and Load Balancing

Scalable Internet Services and Load Balancing Scalable Services and Load Balancing Kai Shen Services brings ubiquitous connection based applications/services accessible to online users through Applications can be designed and launched quickly and

More information

Lizy Kurian John Electrical and Computer Engineering Department, The University of Texas as Austin

Lizy Kurian John Electrical and Computer Engineering Department, The University of Texas as Austin BUS ARCHITECTURES Lizy Kurian John Electrical and Computer Engineering Department, The University of Texas as Austin Keywords: Bus standards, PCI bus, ISA bus, Bus protocols, Serial Buses, USB, IEEE 1394

More information

Achieving Nanosecond Latency Between Applications with IPC Shared Memory Messaging

Achieving Nanosecond Latency Between Applications with IPC Shared Memory Messaging Achieving Nanosecond Latency Between Applications with IPC Shared Memory Messaging In some markets and scenarios where competitive advantage is all about speed, speed is measured in micro- and even nano-seconds.

More information

CDBMS Physical Layer issue: Load Balancing

CDBMS Physical Layer issue: Load Balancing CDBMS Physical Layer issue: Load Balancing Shweta Mongia CSE, School of Engineering G D Goenka University, Sohna Shweta.mongia@gdgoenka.ac.in Shipra Kataria CSE, School of Engineering G D Goenka University,

More information

Enhance Service Delivery and Accelerate Financial Applications with Consolidated Market Data

Enhance Service Delivery and Accelerate Financial Applications with Consolidated Market Data White Paper Enhance Service Delivery and Accelerate Financial Applications with Consolidated Market Data What You Will Learn Financial market technology is advancing at a rapid pace. The integration of

More information

Internet Protocols Fall 2004. Outline

Internet Protocols Fall 2004. Outline Internet Protocols Fall 2004 Lecture 7 Andreas Terzis 9/29/04 Outline Internet Protocol CIDR Forwarding CS 449/Fall 04 2 1 Classless Interdomain Routing (CIDR) Addressing The IP address space is broken

More information

Load Distribution in Large Scale Network Monitoring Infrastructures

Load Distribution in Large Scale Network Monitoring Infrastructures Load Distribution in Large Scale Network Monitoring Infrastructures Josep Sanjuàs-Cuxart, Pere Barlet-Ros, Gianluca Iannaccone, and Josep Solé-Pareta Universitat Politècnica de Catalunya (UPC) {jsanjuas,pbarlet,pareta}@ac.upc.edu

More information

Hardware Configuration Guide

Hardware Configuration Guide Hardware Configuration Guide Contents Contents... 1 Annotation... 1 Factors to consider... 2 Machine Count... 2 Data Size... 2 Data Size Total... 2 Daily Backup Data Size... 2 Unique Data Percentage...

More information

Architectural Level Power Consumption of Network on Chip. Presenter: YUAN Zheng

Architectural Level Power Consumption of Network on Chip. Presenter: YUAN Zheng Architectural Level Power Consumption of Network Presenter: YUAN Zheng Why Architectural Low Power Design? High-speed and large volume communication among different parts on a chip Problem: Power consumption

More information

EFFICIENT DETECTION IN DDOS ATTACK FOR TOPOLOGY GRAPH DEPENDENT PERFORMANCE IN PPM LARGE SCALE IPTRACEBACK

EFFICIENT DETECTION IN DDOS ATTACK FOR TOPOLOGY GRAPH DEPENDENT PERFORMANCE IN PPM LARGE SCALE IPTRACEBACK EFFICIENT DETECTION IN DDOS ATTACK FOR TOPOLOGY GRAPH DEPENDENT PERFORMANCE IN PPM LARGE SCALE IPTRACEBACK S.Abarna 1, R.Padmapriya 2 1 Mphil Scholar, 2 Assistant Professor, Department of Computer Science,

More information

International Journal of Scientific & Engineering Research, Volume 4, Issue 11, November-2013 349 ISSN 2229-5518

International Journal of Scientific & Engineering Research, Volume 4, Issue 11, November-2013 349 ISSN 2229-5518 International Journal of Scientific & Engineering Research, Volume 4, Issue 11, November-2013 349 Load Balancing Heterogeneous Request in DHT-based P2P Systems Mrs. Yogita A. Dalvi Dr. R. Shankar Mr. Atesh

More information

An Adaptive Load Balancing to Provide Quality of Service

An Adaptive Load Balancing to Provide Quality of Service An Adaptive Load Balancing to Provide Quality of Service 1 Zahra Vali, 2 Massoud Reza Hashemi, 3 Neda Moghim *1, Isfahan University of Technology, Isfahan, Iran 2, Isfahan University of Technology, Isfahan,

More information

Benchmarking Cassandra on Violin

Benchmarking Cassandra on Violin Technical White Paper Report Technical Report Benchmarking Cassandra on Violin Accelerating Cassandra Performance and Reducing Read Latency With Violin Memory Flash-based Storage Arrays Version 1.0 Abstract

More information

NetFlow Performance Analysis

NetFlow Performance Analysis NetFlow Performance Analysis Last Updated: May, 2007 The Cisco IOS NetFlow feature set allows for the tracking of individual IP flows as they are received at a Cisco router or switching device. Network

More information

A Content-Based Load Balancing Algorithm for Metadata Servers in Cluster File Systems*

A Content-Based Load Balancing Algorithm for Metadata Servers in Cluster File Systems* A Content-Based Load Balancing Algorithm for Metadata Servers in Cluster File Systems* Junho Jang, Saeyoung Han, Sungyong Park, and Jihoon Yang Department of Computer Science and Interdisciplinary Program

More information

Factors to Consider When Designing a Network

Factors to Consider When Designing a Network Quality of Service Routing for Supporting Multimedia Applications Zheng Wang and Jon Crowcroft Department of Computer Science, University College London Gower Street, London WC1E 6BT, United Kingdom ABSTRACT

More information

Parallel Computing for Option Pricing Based on the Backward Stochastic Differential Equation

Parallel Computing for Option Pricing Based on the Backward Stochastic Differential Equation Parallel Computing for Option Pricing Based on the Backward Stochastic Differential Equation Ying Peng, Bin Gong, Hui Liu, and Yanxin Zhang School of Computer Science and Technology, Shandong University,

More information

MAXIMIZING RESTORABLE THROUGHPUT IN MPLS NETWORKS

MAXIMIZING RESTORABLE THROUGHPUT IN MPLS NETWORKS MAXIMIZING RESTORABLE THROUGHPUT IN MPLS NETWORKS 1 M.LAKSHMI, 2 N.LAKSHMI 1 Assitant Professor, Dept.of.Computer science, MCC college.pattukottai. 2 Research Scholar, Dept.of.Computer science, MCC college.pattukottai.

More information

Network Architecture and Topology

Network Architecture and Topology 1. Introduction 2. Fundamentals and design principles 3. Network architecture and topology 4. Network control and signalling 5. Network components 5.1 links 5.2 switches and routers 6. End systems 7. End-to-end

More information

Dual Mechanism to Detect DDOS Attack Priyanka Dembla, Chander Diwaker 2 1 Research Scholar, 2 Assistant Professor

Dual Mechanism to Detect DDOS Attack Priyanka Dembla, Chander Diwaker 2 1 Research Scholar, 2 Assistant Professor International Association of Scientific Innovation and Research (IASIR) (An Association Unifying the Sciences, Engineering, and Applied Research) International Journal of Engineering, Business and Enterprise

More information

TRILL Large Layer 2 Network Solution

TRILL Large Layer 2 Network Solution TRILL Large Layer 2 Network Solution Contents 1 Network Architecture Requirements of Data Centers in the Cloud Computing Era... 3 2 TRILL Characteristics... 5 3 Huawei TRILL-based Large Layer 2 Network

More information

RSVP- A Fault Tolerant Mechanism in MPLS Networks

RSVP- A Fault Tolerant Mechanism in MPLS Networks RSVP- A Fault Tolerant Mechanism in MPLS Networks S.Ravi Kumar, M.Tech(NN) Assistant Professor Gokul Institute of Technology And Sciences Piridi, Bobbili, Vizianagaram, Andhrapradesh. Abstract: The data

More information

Internetworking and Internet-1. Global Addresses

Internetworking and Internet-1. Global Addresses Internetworking and Internet Global Addresses IP servcie model has two parts Datagram (connectionless) packet delivery model Global addressing scheme awaytoidentifyall H in the internetwork Properties

More information

APPENDIX 1 USER LEVEL IMPLEMENTATION OF PPATPAN IN LINUX SYSTEM

APPENDIX 1 USER LEVEL IMPLEMENTATION OF PPATPAN IN LINUX SYSTEM 152 APPENDIX 1 USER LEVEL IMPLEMENTATION OF PPATPAN IN LINUX SYSTEM A1.1 INTRODUCTION PPATPAN is implemented in a test bed with five Linux system arranged in a multihop topology. The system is implemented

More information

Configuring a Load-Balancing Scheme

Configuring a Load-Balancing Scheme Configuring a Load-Balancing Scheme Last Updated: October 5, 2011 This module contains information about Cisco Express Forwarding and describes the tasks for configuring a load-balancing scheme for Cisco

More information

Storage Area Network Design Overview Using Brocade DCX 8510. Backbone Switches

Storage Area Network Design Overview Using Brocade DCX 8510. Backbone Switches Storage Area Network Design Overview Using Brocade DCX 8510 Backbone Switches East Carolina University Paola Stone Martinez April, 2015 Abstract The design of a Storage Area Networks is a very complex

More information

Implementation of a Lightweight Service Advertisement and Discovery Protocol for Mobile Ad hoc Networks

Implementation of a Lightweight Service Advertisement and Discovery Protocol for Mobile Ad hoc Networks Implementation of a Lightweight Advertisement and Discovery Protocol for Mobile Ad hoc Networks Wenbin Ma * Department of Electrical and Computer Engineering 19 Memorial Drive West, Lehigh University Bethlehem,

More information

Towards TCAM-based Scalable Virtual Routers

Towards TCAM-based Scalable Virtual Routers Towards -based Scalable Virtual Routers Layong Luo *, Gaogang Xie *, Steve Uhlig, Laurent Mathy, Kavé Salamatian, and Yingke Xie * * Institute of Computing Technology, Chinese Academy of Sciences (CAS),

More information

Multi-layer Structure of Data Center Based on Steiner Triple System

Multi-layer Structure of Data Center Based on Steiner Triple System Journal of Computational Information Systems 9: 11 (2013) 4371 4378 Available at http://www.jofcis.com Multi-layer Structure of Data Center Based on Steiner Triple System Jianfei ZHANG 1, Zhiyi FANG 1,

More information

Quality of Service using Traffic Engineering over MPLS: An Analysis. Praveen Bhaniramka, Wei Sun, Raj Jain

Quality of Service using Traffic Engineering over MPLS: An Analysis. Praveen Bhaniramka, Wei Sun, Raj Jain Praveen Bhaniramka, Wei Sun, Raj Jain Department of Computer and Information Science The Ohio State University 201 Neil Ave, DL39 Columbus, OH 43210 USA Telephone Number: +1 614-292-3989 FAX number: +1

More information

Entropy-Based Collaborative Detection of DDoS Attacks on Community Networks

Entropy-Based Collaborative Detection of DDoS Attacks on Community Networks Entropy-Based Collaborative Detection of DDoS Attacks on Community Networks Krishnamoorthy.D 1, Dr.S.Thirunirai Senthil, Ph.D 2 1 PG student of M.Tech Computer Science and Engineering, PRIST University,

More information

An Energy Efficient Location Service for Mobile Ad Hoc Networks

An Energy Efficient Location Service for Mobile Ad Hoc Networks An Energ Efficient Location Service for Mobile Ad Hoc Networks Zijian Wang 1, Euphan Bulut 1 and Boleslaw K. Szmanski 1, 1 Department of Computer Science, Rensselaer Poltechnic Institute, Tro, NY 12180

More information

Performance Evaluation of AODV, OLSR Routing Protocol in VOIP Over Ad Hoc

Performance Evaluation of AODV, OLSR Routing Protocol in VOIP Over Ad Hoc (International Journal of Computer Science & Management Studies) Vol. 17, Issue 01 Performance Evaluation of AODV, OLSR Routing Protocol in VOIP Over Ad Hoc Dr. Khalid Hamid Bilal Khartoum, Sudan dr.khalidbilal@hotmail.com

More information

Original-page small file oriented EXT3 file storage system

Original-page small file oriented EXT3 file storage system Original-page small file oriented EXT3 file storage system Zhang Weizhe, Hui He, Zhang Qizhen School of Computer Science and Technology, Harbin Institute of Technology, Harbin E-mail: wzzhang@hit.edu.cn

More information

Bit Vector Algorithms Enabling High-Speed and Memory- Efficient Firewall Blacklisting

Bit Vector Algorithms Enabling High-Speed and Memory- Efficient Firewall Blacklisting Bit Vector Algorithms Enabling High-Speed and Memory- Efficient Firewall Blacklisting Lane Thames Randal Abler David Keeling Georgia Institute of Technology Georgia Institute of Technology Georgia Institute

More information