Intelligent Application Switch Supporting TCP

Recent programmable switches allow developers to profoundly optimize network elements. In this paper, we introduce a concept of an application switch that supports a network application based on programmable switches. A developer can optimize not only server computers but also network switches in order to improve application performance by implementing part of functions of the network application with an application switch. We then propose a method to apply an application switch for a database management system (DBMS) using TCP. The method migrates a TCP connection from a server computer to the application switch when the switch replies a query. The method reversely migrates it to the server computer again when a query is processed by the server. The application switch manages the TCP sequence number and the Ack number for these migrations. We constructed key-value store (KVS), which is one of DBMSs, using Cassandra and evaluated our method. Our evaluation showed that the proposed method improved the performance by 69% at most.

[1]  Saneyasu Yamaguchi,et al.  Improving Dynamic Scaling Performance of Cassandra , 2017, IEICE Trans. Inf. Syst..

[2]  Masato Oguchi,et al.  iSCSI analysis system and performance improvement of sequential access in a long‐latency environment , 2006 .

[3]  C. Partridge,et al.  Computing the internet checksum , 1989, CCRV.

[4]  Masato Oguchi,et al.  Trace system of iSCSI storage access , 2005, The 2005 Symposium on Applications and the Internet.

[5]  EDDIE KOHLER,et al.  The click modular router , 2000, TOCS.

[6]  Akihiro Nakao,et al.  Application performance improvement with application aware DPN switches , 2016, 2016 18th Asia-Pacific Network Operations and Management Symposium (APNOMS).

[7]  Akihiro Nakao,et al.  Network Application Performance Improvement with Deeply Programmable Switch , 2016, MobiQuitous.

[8]  Machiko Toyoda,et al.  TCP Congestion Window Control on an ISCSI Read Access in a Long-Latency Environment , 2005, Communication Systems and Applications.

[9]  George Varghese,et al.  P4: programming protocol-independent packet processors , 2013, CCRV.

[10]  Akihiro Nakao,et al.  Application switch using DPN for improving TCP based data center applications , 2017, 2017 IFIP/IEEE Symposium on Integrated Network and Service Management (IM).

[11]  Anil Rijsinghani,et al.  Computation of the Internet Checksum via Incremental Update , 1994, RFC.

[12]  Saneyasu Yamaguchi,et al.  Improving RTT Fairness on CUBIC TCP , 2013, 2013 First International Symposium on Computing and Networking.