Title
SD-P2: a Novel Software Defined Public Internet Infrastructure for Joint Cloud Computing
Abstract
A key challenge for addressing the cross-cloud issues like a burst of load, restricted data processing due to regulation, and platform lock-in is to rapidly construct transparent wide-area network infrastructure among different cloud vendors in the face of variable traffic demands and failures. In this paper we propose Software Defined Public Internet Infrastructure (SD-P2), a new elastic and resilient architecture for Joint Cloud Computing, where we decouple the existing design of network functions into data processing components to perform flexible packet processing inside the network and network scheduling components to provide fair, efficient and dynamic resource scheduling capabilities on SD-P2. Our data processing instances are architected around efficient pipelines utilizing 4-layer VPN for security transmission and DPDK for high performance network I/O, and our network scheduling and service chain management are designed by complementing the capabilities of the network scheduler based on the optimal vNF placement model. Our evaluation shows that, compared to existing cross-cloud solutions, (1) we provide the proportional fairness share of wide-area heterogeneous network resource and significantly improve network performance (throughput and latency) by optimally and dynamically placing virtualized network functions across several clouds, (2) we are able to reach a throughput of 10Gbit/sec, with an added latency overhead of between 10ms, (3) we are able to reduce the communication costs by 83% compared to traditional vNF placement solutions.
Year
DOI
Venue
2019
10.1109/SOSE.2019.00052
2019 IEEE International Conference on Service-Oriented System Engineering (SOSE)
Keywords
Field
DocType
Cloud computing,IP networks,Dynamic scheduling,Computer architecture,Noise measurement,Virtual private networks
Computer science,Scheduling (computing),Network scheduler,Packet processing,Throughput,Heterogeneous network,Dynamic priority scheduling,Cloud computing,Network performance,Distributed computing
Conference
ISBN
Citations 
PageRank 
978-1-7281-1442-2
0
0.34
References 
Authors
0
3
Name
Order
Citations
PageRank
Daochao Huang1696.47
Shijun Shen200.34
Xinran Liu32813.23