- Tuesday, 13 June, 11:30-13:00, Room Theatre Piki Hall
- Session Chair: Zhisheng Niu (Tsinghua University, P.R. China)
Jang-Ping Sheu, Po-Yao Wang and Jagadeesha Rb (National Tsing Hua University, Taiwan)
TCAM capacity problem is an important issue in Software-Defined Networking. Rule caching is an efficient technique to solve the TCAM capacity problem. However, there exists rule dependency problem in wildcard-rule caching technique. In this paper, we utilize cover-set method to solve rule dependency problem and propose a wildcard-rule caching algorithm to cache rules into TCAM. In addition, we also propose a cache replacement algorithm considering both temporal and spatial localities. According to the simulation results, our wildcard-rule caching algorithm and cache replacement algorithm have better performance than previous works in terms of caching ratio and hit ratio, respectively.
Mael Kimmerlin (Aalto University, Finland); Peer Hasselmeyer (NEC Laboratories Europe, NEC Europe Ltd., Germany); Seppo Heikkilä (Helsinki Institute of Physics, Switzerland); Max Plauth (Hasso Plattner Institute, Germany); Paweł Parol (Orange Polska, Poland); Pasi Sarolahti (Aalto University, Finland)
Cloud federation is receiving increasing attention due to the benefits of resilience and locality it brings to cloud providers and users. Our analysis of three diverse use cases shows that existing solutions are not addressing the federation needs of such use case applications. In this paper, we present an alternative approach to networking federation, providing a model based on cloud-to-cloud agreements. In our scenarios, companies hosting their own OpenStack clouds need to run machines transparently in another cloud, provided by a company they have an agreement with. Our solution provides multiple benefits for cloud providers and users detailed in this paper. Our implementation outperforms the VPNaaS solution in OpenStack in terms of throughput.
Eduard Grasa (Fundació i2CAT, Internet i Innovació Digital a Catalunya, Spain); Leonardo Bergesio (i2CAT, Spain); Miquel Tarzan-Lorente (i2Cat Foundation, Spain); Diego Lopez (Telefonica I+D, Spain); John Day and Lou Chitkushev (Boston University, USA)
Network renumbering in the IP world is a complicated and expensive procedure that has to be carefully planned and executed to avoid routing, security (firewall, ACLs) and transport connection integrity problems. The source of most of these issues is in the lack of a complete naming and addressing architecture in the TCP/IP protocol suite. This paper analyses the issues related to IP networks renumbering, identifying its root causes. Then it looks into how these issues affect renumbering in networks based on RINA, a network architecture with a complete naming scheme. Theoretical analysis backed up by experimentation results indicate that renumbering in RINA networks not only is seamless (can be done without impacting existing flows) but also does not require any special mechanisms.
Shireesh Bhat, Robinson Udechukwu, Rudra Dutta and George N. Rouskas (North Carolina State University, USA)
Network service composition provided as a service in an Open Marketplace enables users to obtain customized end-to-end composed service(s) using the services advertised by the providers in the marketplace. By providing a semantic language for advertising services and offering choice for the composed service(s) we provide a level playing field for the providers and alternatives for the users to choose from based on their requirement. This is similar to the services offered in the cloud, but without the provider monopoly or the limitation of having to select from limited options.
Elisa Jimeno (Atos); Javier Garcia Lloreda (Atos, Spain); Pouria Sayyad Khodashenas (i2CAT Foundation (i2CAT), Spain); August Betzler (i2CAT Foundation, Spain); Irena Trajkovska and Luca Del Vecchio (ZHAW, Switzerland); Alan Whitehead (Ip Access Ltd, United Kingdom (Great Britain))
To meet the communication expectations of the future, mobile networks need to evolve quicker than ever toward systems with ultra-low latency, huge traffic volume and higher data rates. Cloud-enabled radio systems have been introduced as a promising solution to meet these demands with the help of network functions virtualization techniques and network edge processing capabilities, which allow for increased resource pooling, scalability, layer interworking and spectral efficiency. Despite their potential benefits, joint radio-cloud systems pose technical challenges on the network management and orchestration, especially on ensuring the Quality of Service (QoS). To this end, having a complete loop of monitoring, decision-making and reaction is essential. However, considering the fact that the radio and the cloud parameters are inherently disparate, forming such a loop in a joint radio-cloud environment is very challenging. This challenge becomes more difficult in multi-tenant (operator) scenarios, targeted by 5G, where ensuring the QoS for one tenant should not violate the QoS of the others. This paper intends to state the problem from a 5GPPP H2020 perspective, and discusses a possible solution within the context of the SESAME project.