High-Performance IPv6 Lookup with Real-Time Updates Using Hierarchical-Balanced Search Tree

SegMobaTree: The Segmented Multilayer Online Balanced Tree for high-performance IPv6 Lookup in the Edge Network

“…The binary-based algorithms: The binary-based algotihm Hi-BST [7] applies Treap to achieve high update speed and low memory cost, but limited lookup speed with too many memory accesses. Furthermore, its fixed scheme to split segments of prefix lengths is only designed for backbone IPv6 rulesets, the edge IPv6 rulesets requires the dynamic scheme to find suitable segments.…”

“…Some trie-based algorithms like LC-Tries [3], Tree-bitmap [4], SAIL [5], and Poptrie [6] have disadvantages in IPv6 rulesets with longer prefix lengths. The binary-based algorithm Hi-BST [7] can perform high-speed update with low memory cost, but its lookup speed is limited beacuse the inefficient data structure produces too many memory accesses. Therefore, the existing algorithms are insufficient to support current IPv6 lookup in the edge network.…”

Scalable Balanced Pipelined IPv6 Lookup Algorithm

“…Most of the lookup algorithms are designed to support IPv4 addresses, but they usually do not scale well to longer IPv6 addresses. There is an ongoing effort to develop IP lookup algorithms that achieve efficient IPv6 addresses support [2], [7], [9]- [10], [12], [14], [16], [19]- [20], [22], [25]- [26], [29]. All lookup algorithms can be classified into three main categories [2]: tree based algorithms [3]- [22], Ternary Content-Addressable Memory (TCAM) based algorithms [23]- [25], and hash based algorithms [26]- [29].…”

“…The priority tree technique can be applied to m-ry trees as well [8]. In [16], a hierarchical-balanced search tree is constructed. This tree contains only non-empty nodes that store prefix range information.…”

“…We compare our proposed lookup algorithm to several recently proposed tree based lookup algorithms. We select Linear Pipelined IPv6 Lookup Architecture (LPILA) [14], Hierarchical Balanced Search Tree (Hi-BST) [16], Splitting Approach to IP Lookup with Population Counting (SAIL-PC) [18], and Multilevel Length-based-classified Index Table (MLIT) [19] algorithms because each of them uses some of the techniques used in BPL (balancing technique, pipeline technique, and only non-empty nodes storage). To achieve high lookup throughput, the on-chip memories need to be used for the lookup table implementation.…”