Tutorials
Abbas Jamalipour
University of Sydney, AustraliaTitle: Scaling Dense-Traffic Cellular Networks Through Software Defined Networking
Abstract The exponential growth of cellular network traffic and flattening spectral efficiency are two of the biggest challenges facing 4G and future 5G wireless cellular networks. On the demand side, network traffic is exponentially growing and becoming much more diverse, while on the supply side the available spectrum and spectral efficiency are flattening out. The common approach to overcome the problem of increasing per-user capacity and spectral efficiency is to make cells smaller and bring base stations closer to mobile hosts, resulting in extremely dense network deployments. The current architectures are however unable to support such extremely dense cellular deployments. Adopting fundamentals of software defined networking (SDN) in designing layers of system and abstractions can simplify the deployment and management of extremely dense wireless cellular networks. SDN can programmatically decouple network architectures from infrastructure, thus making it easier for deploying new applications and services, as well as for tuning network policy and performance. SDN enables distribution of data-plane rules over multiple, low-cost network switches, hence reducing the scalability of the packet gateway and enabling flexible handling of data traffic over the cellular network. This tutorial provides a comprehensive review of current and future SDN-based approach to solve these problems.
Noriaki Kamiyama
Osaka University & NTT Network Technology Laboratories, JapanTitle: Advances in Reducing Web Response Time
Abstract In recent years, web surfing is one of the most common and popular service provided on the Internet. The ratio of dynamic objects generated when executing programs, e.g., JSP and Ajax has increased, and the servers providing each object have been diversified, e.g., objects of advertisements are obtained from dedicated servers. Therefore, the objects that construct one website have been diversified, and the complexity of communication pattern generated at web browsing has continued to increase. As a result, the web response time, i.e., the waiting time until webpages are displaied at browser, has been increased. For example, when browsing about 1,000 most popular webpages from Tokyo, the response time of 50% webpages was more than 4 seconds. This tutorial gives an overview of various advanced techniques accelerating web response time. First, I abstruct the recent web technologies including the fundamental bodies, e.g., HTML and HTTP, and the recent techniques generating dynamic objects, e.g., Ajax, Servlet, and JSP. Next, I give an overview of some recent works investigating what happens in web browsing service, e.g., analyzing the deployment pattern of web objects. Finally, I explain some popular techniques reducing the web response time including CDN and SPDY, and Inlining, as well as the recent ideas reducing the web response time, e.g., Inlining, edge/fog computing, and prefetching.
Dimitri Papadimitriou
Nokia Bell Labs, BelgiumTitle: New Challenges in Network Optimization
Abstract The recent evolution of communication networks has induced new challenges in mixed-integer network optimization ranging from problem formulation to computational methods. This tutorial covers in detail five of these problems i) the progressive unification of communication networks with information systems (involving storage and processing) leads to combined routing design - capacitated facility/hub location models where digital objects replace physical goods, introduction of delay-awareness and other temporal constraints (waiting time, replenishment time, etc.), ii) in the same context, the design of resilient network infrastructure can't be designed anymore independently of end-point (content servers, caches) properties and information spatio-temporal distribution properties leading to re-consider classical network resource optimization problems, iii) the expansion of communication networks to heterogeneous environments and running conditions requires incorporating uncertainty in demands and topology parameters in order to formulate robust counterpart of combined network and routing design problems, iv) virtual data centers enable flexible allocation of capacity to serve customer demands by aggregating physical resources taken out of distributed resource pools accessible by customers via dedicated gateways, the corresponding optimization problem can be modeled by combining the three-level capacitated facility location with a flow routing problem, and v) finally, and for more challenging, as optics added a spectral dimension to spatial routing, information adds a semantic dimension to spatial routing (both dimensions having large impact on scale), leading in turn to introduce categorical constraints. In the second part of this tutorial we detail their implications in terms of computational methods and techniques when solving large-scale problems.