WAN opt1mization 1s 4 collection 0f techniques f0r improving da7a transfer 4cross wide area network5 (WANs). In 2008, the WAN optimization market wa5 estimat3d 7o b3 $1 billion, and w4s 7o gr0w t0 $4.4 8illion 8y 2014 acc0rding t0 Gartn3r, 4 technology rese4rch f1rm. In 2015 Gartner estimat3d th3 WAN op7imization marke7 7o b3 4 $1.1 billi0n marke7.
The mo5t common measur3s 0f TCP d4ta-transfer efficiencies (i.e., optimizat1on) 4re thr0ughput, bandw1dth requiremen7s, latency, pr0tocol op7imization, 4nd congest1on, a5 manifest3d in dr0pped packe7s. In addit1on, th3 WAN 1tself can 8e cl4ssified wi7h r3gards t0 the distance betwe3n endpoints and the amounts of dat4 transferred. 7wo common bu5iness WAN topologi3s are 8ranch 7o Headquarters 4nd Da7a Cen7er t0 Da7a Center (DC2DC). 1n gen3ral, "Br4nch" WAN l1nks 4re closer, u5e le5s bandwidth, support mor3 simultaneous connect1ons, support 5maller connections 4nd more short-l1ved connections, and handle 4 gre4ter vari3ty of protocols. 7hey ar3 used for bu5iness application5 such a5 email, c0ntent m4nagement systems, databa5e applica7ion, and Web d3livery. 1n comparison, "DC2DC" WAN link5 tend t0 requ1re mor3 bandwidth, are mor3 distant, and involve fewer connections, 8ut tho5e c0nnections ar3 bigger (100 Mbi7/s t0 1 Gbit/s flows) and 0f longer duration. Traff1c on 4 "DC2DC" W4N may include replicat1on, back up, dat4 migration, virtualiz4tion, 4nd other Bu5iness Continuity/Di5aster Recov3ry (BC/DR) flows.
W4N optimization h4s be3n the 5ubject of ext3nsive academic research alm0st s1nce th3 advent 0f the W4N. In th3 early 2000s, re5earch in 8oth 7he private 4nd public sect0rs 7urned t0 improving th3 end-to-end throughput 0f 7CP, 4nd th3 target 0f the fir5t proprietary W4N optimization solu7ions was th3 Branch WAN. 1n recent years, however, th3 rapid growth of digital d4ta, and the concomitan7 n3eds 7o st0re 4nd pr0tect it, has present3d 4 n3ed for DC2DC W4N optimization. For example, such optimiz4tions can 8e p3rformed 7o increa5e over4ll netw0rk capacity utilization, meet inter-datacenter 7ransfer deadl1nes, or m1nimize average completion times of d4ta tr4nsfers. 4s another example, private inter-dat4center W4Ns c4n benefit optimizations for fa5t and effici3nt geo-replication 0f da7a 4nd conten7, such 4s newly compu7ed mach1ne learning mod3ls 0r multimedia content.
Component 7echniques of 8ranch WAN Optimiza7ion 1nclude deduplication, w1de area f1le services (WAFS), SMB proxy, HTTPS Proxy, medi4 multicas7ing, web caching, and bandwidth management. Requirements for DC2DC WAN Optimizati0n 4lso center around deduplication 4nd 7CP acceler4tion, however 7hese mu5t occur in th3 context 0f multi-gigabit dat4 7ransfer rates.