László Gyarmati

Scafida: a scale-free network inspired data center architecture

Data centers have a crucial role in current Internet architecture supporting content-centric networking. State-of-the-art data centers have different architectures like fat-tree, DCell, or BCube. However, their architectures share a common property: symmetry. Due to their symmetric nature, a tricky point with these architectures is that they are hard to be extended in small quantities. Contrary to state-of-the-art data center architectures, we propose an asymmetric data center topology generation method called Scafida inspired by scale-free networks; these data centers have not only small diameters and high fault tolerance, inherited by scale-free networks, but can also be scaled in smaller and less homogenous increments. We extend the original scale-free network generation algorithm of Barabasi and Albert to meet the physical constraints of switches and routers. Despite the fact that our method artificially limits the node degrees in the network, our data center architectures keep the preferable properties of scale-free networks. Based on extensive simulations we present preliminary results that are promising regarding the error tolerance, scalability, and flexibility of the architecture.

How can architecture help to reduce energy consumption in data center networking

Reducing the power consumption of data centers has recently been received much attention from the research community and the industry alike. It is because the benefits are manifold: for the data centers themselves and towards the environment. A number of issues concerning with the consumption of the servers or the switches have been investigated. However, the energy requirement of the data center architecture has not yet been sufficiently addressed. In this paper, we address this issue and make two key contributions. Before delving into the details of possible solutions for reducing power of existing data center architectures, we first characterize the impact of architectural parameters on the power consumption of data centers by presenting preliminary results on the power consumption of state-of-the-art data center structures, namely BCube [8], DCell [9], and Fat-tree [1, 11, 7]. Secondly, based on the insights and lessons learnt from the analysis, we present our vision on possible practical solutions to reduce the energy usage of data center networking.

Multidomain shared protection with limited information via MPP and p-cycles

The Internet consists of a collection of more than 21,000 domains called autonomous systems operated mostly under different authorities (operators-providers) that, although they cooperate over different geographical areas, compete in a country or other area. Recently, the path computation element concept has been proposed to generalized multiprotocol label switching controlled optical borne networks to make routing decisions for interdomain connections taking into account traffic engineering, quality of service, and resilience considerations. Still the question of protection shareability emerges. For dedicated protection it is enough to know the topology of the network to be able to calculate disjoint paths. However, to reduce network resource usage by sharing of protection resources (e.g., end-to-end shared protection) it is also mandatory to know the exact working and protection path pairs for all the demands. This can be checked within a domain where not only the full topology and link-state information is flooded but also the working and protection paths are known for each connection; however; over the domain boundaries for security and scalability reasons no such information is being spread. We propose using two techniques that do not require flooding the information on working and protection paths while still allowing the sharing of resources. These two techniques are the multidomain p-cycles and the multidomain multipath routing with protection. After explaining the principles of these methods we evaluate the trade-off between the resource requirement and availability of these techniques by simulations.

MPP: Optimal Multi-Path Routing with Protection

Multi-path routing routes the upcoming demand over multiple paths. This is feasible in networks where the demand can be split into multiple parallel flows, i.e., the network is capable to perform inverse multiplexing (e.g., ngSDH and OTN via VCat) and where the connectivity of the topology also allows it. In this paper we propose the MPP (Multi-Path Protection) scheme, where the demand is not only routed, but also protected using multiple parallel paths. We give a Linear Programming (LP) formulation of the problem that finds the globally optimal solution. The main features of our approach are twofold. First, it always finds a trade-off between the number of parallel paths to be used and the total length of these paths while it balances the traffic optimally between them. Second, we use LP with real variables that is solvable in polynomial time instead of Integer LP (ILP) and still we are able to avoid branching of the flows in nodes different than the source and the target of the considered demand.

Multi-Domain Resilience: Can I Share Protection Resources with my Competitors?

The Internet consists of a collection of more than 21000 domains called Autonomous Systems (AS) operated mostly under different authorities (operators/providers) that although co-operate over different geographical areas, they compete in a country or other area. Today BGP is the de facto standard for exchanging reachability information over the domain boundaries and for inter-domain routing. The GMPLS controlled optical beared networks are expected to have similar architecture, however, more information has to be carried for TE, resilience and QoS purposes. Therefore, extensions of BGP and of PNNI as well as the PCE have been proposed. Still in all cases emerges the question of protection shareability. For dedicated protection it is enough to know the topology of the network to be able to calculate disjoint paths. However, to be able to perform sharing of protection resources (shared protection) it is not enough to know the topology, but it is mandatory to know exact working and protection path pairs for all the demands, since protection paths can share a certain resource only if there is no such a pair of working paths that contain any element from the same Shared Risk Group (SRG). This can be checked within a domain where the full topology and link-state information is flooded, however, over the domain boundaries for security and scalability reasons no such information is being spread. In this paper we propose using two techniques that do not require flooding the information on working and protection paths while they still allow sharing of resources. These two techniques are the Multi-Domain p- Cycles (MD-PC) and the Multi-Domain Multi-Path Routing with Protection (MD-MPP). After explaining the principles of these methods we give illustrative results.

Migrating to IPv6: A game-theoretic perspective

The rate of deployment and adoption issues of new network technologies, IPv6 in particular, have recently been hotly debated in the research community. However, the question of how protocols migrate, especially the dynamics of migration, to new paradigms is still largely open. In this paper, we address the issue from a game theoretic point of view. We model and analyze the profit maximizing strategies of Autonomous Systems (ASes); both the properties of ASes and the topology of the Internet is considered. The contribution of our work is threefold. First, we propose an economic model of the ASes and their relations from the IPv4-IPv6 migration viewpoint. Second, we apply the findings of evolutionary dynamics on the problem of migration by incorporating Internet-specific properties to the evolutionary model, namely the size of the ASes and the cost of migration. The analyses show that even if IPv6 has higher payoff than IPv4, the whole migration does not happen always fast. Finally, extensive simulations are carried out based on the proposed models to illustrate the impacts of different parameters on the IPv6 migration dynamics in realistic scenarios.

Energy Efficiency of Data Centers

As the power consumption has a significant and continuously increasing part of the operational expenses of data centers, energy efficient data center networking has received special attention from the academic and industrial research community recently. The complex design of data centers provides several directions toward more energy efficiency, including the consumption of servers and network equipments. We address the issue of power consumption of data centers from a higher-level point of view by analyzing the energy efficiency of data center architectures. We review the state-of-the-art data center architectures including BCube, DCell, fat-tree, and Scafida and present evaluate their energy-efficiency quantitatively. In addition, the trade-off between the power consumption and the performance of the data center is investigated. Next, other aspects of data centers’ energy efficiency is revealed including thermal control techniques, energy management systems. Finally, a standard data center energy efficiency metric called PUE (Power Usage Effectiveness) is presented.

Understanding the impact of loyal user behaviour on Internet access pricing: a game-theoretic framework

In this paper, we investigate the impacts of user behaviour—user loyalty in particular—on pricing strategies of Internet Service Providers (ISPs) for a profitable yet sustainable Internet access marketplace. We carry out an extensive empirical analysis of customer loyalty issues of ISP markets including our own survey in the Hungarian ISP market. Based on the empirical results, we propose a realistic user loyalty model, the price difference dependent loyalty model. Next, we apply the loyalty model in game-theoretical framework where optimal Internet access pricing strategies are expressed. Our game-theoretic framework includes both short-term and long-term interaction cases (single-shot and repeated games, respectively) and is capable of dealing with uncertain as well as dynamic scenarios (Bayesian and Stackelberg games, respectively). Finally, we present the impacts of user loyalty on the prices and profits of ISPs in different scenarios based on simulation results.

SRLG-Disjoint Multi-Path Protection: When LP meets ILP

Multi-Path Routing routes the upcoming demand over multiple paths. This is feasible in networks where the demand can be split into multiple parallel flows, i.e., the network is capable to perform inverse multiplexing (e.g., ngSDH and OTN via VCat) and where the connectivity of the topology also allows it.