Scalable and Adaptive Software Defined Network Management for Cloud-hosted Group Communication Applications

被引:0
作者
Patil, Prithviraj [1 ,4 ]
Hakiri, Akram [2 ]
Shekhar, Shashank [3 ]
Gokhale, Aniruddha [3 ]
机构
[1] MathWorks Inc, Natick, MA 01760 USA
[2] Univ Carthage, ISSAT, Mateur, Bizerte, Tunisia
[3] Vanderbilt Univ, Dept EECS, 221 Kirkland Hall, Nashville, TN 37235 USA
[4] Vanderbilt Univ, 221 Kirkland Hall, Nashville, TN 37235 USA
来源
PROCEEDINGS OF THE 10TH INTERNATIONAL CONFERENCE ON UTILITY AND CLOUD COMPUTING (UCC' 17) | 2017年
关键词
Cloud Computing; Adaptive Multicast; Software Defined Network; OpenFlow; Data Center Networks; MULTICAST; EFFICIENT;
D O I
10.1145/3147213.3147220
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
Group communications form the primary communication pattern for many cloud-hosted applications and cloud infrastructure management services, such as system health monitoring, multimedia distribution, collaborative applications and distributed databases. Although IP multicast has been used to support group communication semantics in diverse Internet-based distributed applications, its deployment in cloud Data Center Networks (DCNs) has been limited due to its higher resource consumption, scalability, and stability issues, which in turn degrades the utility of the cloud. Software Defined Networking (SDN) has enabled the re-engineering of multicast capabilities to overcome these limitations. To that end, this paper presents an autonomous, dynamic and flexible middleware solution called SDN-based Multicast (SDMC), which provides both network load-aware and switch memory-efficient group communication semantics in DCNs. Thus, SDMC improves DCN resource utilization while allowing applications to remain agnostic to the underlying group communication semantics by efficiently toggling between unicast and multicast in accordance with changing network bandwidth and switch memory usage. Empirical studies comparing SDMC with traditional IP multicast shows up to 60% better latency performance for different DCNs topologies, and up to 50% better performance in the switch memory utilization for multicast groups exceeding size 30.
引用
收藏
页码:111 / 120
页数:10
相关论文
共 36 条
[1]  
[Anonymous], 2013, MULTICAST DATA CTR O
[2]  
[Anonymous], 2015, P IEEE GLOBECOM 15
[3]  
[Anonymous], 2012, 2012 4 INT C COMM SY
[4]  
[Anonymous], 2015, 7348 RFC
[5]   Energy-aware resource allocation heuristics for efficient management of data centers for Cloud computing [J].
Beloglazov, Anton ;
Abawajy, Jemal ;
Buyya, Rajkumar .
FUTURE GENERATION COMPUTER SYSTEMS-THE INTERNATIONAL JOURNAL OF ESCIENCE, 2012, 28 (05) :755-768
[6]  
Fan L, 2011, 2011 SECOND INTERNATIONAL CONFERENCE ON EDUCATION AND SPORTS EDUCATION (ESE 2011), VOL III, P264
[7]   On the Definition of Role Mining [J].
Frank, Mario ;
Buhmann, Joachim M. ;
Basin, David .
SACMAT 2010: PROCEEDINGS OF THE 15TH ACM SYMPOSIUM ON ACCESS CONTROL MODELS AND TECHNOLOGIES, 2010, :35-44
[8]   Exploring Server Redundancy in Nonblocking Multicast Data Center Networks [J].
Guo, Zhiyang ;
Yang, Yuanyuan .
IEEE TRANSACTIONS ON COMPUTERS, 2015, 64 (07) :1912-1926
[9]  
Holbrook Hugh, 2015, 4604 RFC
[10]  
Hong Linfeng, 2013, 2012 17 OPT COMM C