An SDN Architecture for Improving Throughput of Large Flows Using Multipath TCP

The recent developments in data center topologies offer high bandwidth capacity and multiple paths between processing nodes. The increased use of cloud computing applications imposes challenges on underlying network infrastructure. To meet these challenges we need efficient traffic management that can maximize the utilization of multi-path data center networks. Path diversity not only provides the opportunity of performing load balancing and improve fault tolerance, but also can be exploited to allocate more bandwidth for large flows and improve the performance of network-limited applications. In this paper, we present an architecture based on Multipath TCP (MPTCP) and Software Defined Networking (SDN) to provide better bandwidth allocation for large flows. The SDN controller, which has global knowledge of network topology and traffic measurements, is capable of making educated routing decisions. The proposed SDN architecture enables applications to achieve better throughput for large flows by initiating new MPTCP subflows. To provide this capability, we modify the Linux kernel MPTCP implementation to enable applications to create additional MPTCP subflows on-demand. These MPTCP subflows are placed on least-congested paths by the centralized controller. The evaluation results obtained from running experiments on the GENI testbed environment show a significant improvement in the throughput of large flows.

[1]  Keke Gai,et al.  Dynamic energy-aware cloudlet-based mobile cloud computing model for green computing , 2016, J. Netw. Comput. Appl..

[2]  Mark Handley,et al.  Improving datacenter performance and robustness with multipath TCP , 2011, SIGCOMM 2011.

[3]  Martín Casado,et al.  The Design and Implementation of Open vSwitch , 2015, NSDI.

[4]  Keke Gai,et al.  Blend Arithmetic Operations on Tensor-Based Fully Homomorphic Encryption Over Real Numbers , 2018, IEEE Transactions on Industrial Informatics.

[5]  Ian Wakeman,et al.  MMPTCP: A multipath transport protocol for data centers , 2016, IEEE INFOCOM 2016 - The 35th Annual IEEE International Conference on Computer Communications.

[6]  Keke Gai,et al.  Energy-aware task assignment for mobile cyber-enabled applications in heterogeneous cloud computing , 2018, J. Parallel Distributed Comput..

[7]  Mark Handley,et al.  Design, Implementation and Evaluation of Congestion Control for Multipath TCP , 2011, NSDI.

[8]  Mark Handley,et al.  TCP Extensions for Multipath Operation with Multiple Addresses , 2020, RFC.

[9]  Charles Clos,et al.  A study of non-blocking switching networks , 1953 .

[10]  Haitao Wu,et al.  BCube: a high performance, server-centric network architecture for modular data centers , 2009, SIGCOMM '09.

[11]  Alan Silva,et al.  On the Benefits of Using Multipath TCP and Openflow in Shared Bottlenecks , 2015, 2015 IEEE 29th International Conference on Advanced Information Networking and Applications.

[12]  David A. Maltz,et al.  Network traffic characteristics of data centers in the wild , 2010, IMC '10.

[13]  Mark Handley,et al.  TCP Extensions for Multipath Operation with Multiple Addresses , 2011 .

[14]  Olivier Bonaventure,et al.  An enhanced socket API for Multipath TCP , 2016, ANRW '16.

[15]  Ankit Singla,et al.  Jellyfish: Networking Data Centers Randomly , 2011, NSDI.

[16]  Amin Vahdat,et al.  A scalable, commodity data center network architecture , 2008, SIGCOMM '08.

[17]  Savvas Zannettou,et al.  Exploiting path diversity in datacenters using MPTCP-aware SDN , 2015, 2016 IEEE Symposium on Computers and Communication (ISCC).