Jason Min Wang

Intra-AS cooperative caching for content-centric networks

The default caching scheme in CCN results in a high redundancy along the symmetric request-response path, and makes the caching system inefficient. Since it was first proposed, much work has been done to improve the general caching performance of CCN. Most new caching schemes attempt to reduce the on-path redundancy by passing information on content redundancy and popularity between nodes. In this paper, we tackle the problem from a different perspective. Instead of curbing the redundancy through special caching decisions in the beginning, we take an orthogonal approach by pro-actively eliminating redundancy via an independent intra-AS procedure. We propose an \textit{intra-AS cache cooperation} scheme, to effectively control the redundancy level within the AS and allow neighbour nodes in an AS to collaborate in serving each others requests. We show via trace-driven simulation, that intra-AS cache cooperation improves the system caching performance and reduces considerably the traffic load on the AS gateway links, which is very appealing from an ISPs perspective.

Progressive caching in CCN

The current Internet is based on the old concept of data networking whose role was to mainly build a fault-tolerant network infrastructure that enables efficient sharing of scarce and expensive resources such as storage and CPU time. Its main feature is the end-to-end principle of communication between any pair of hosts under the client-server model, which was well suited for early Internet applications, e.g., remote login (CPU sharing) and ftp (storage). In recent years, the main usage of the Internet has evolved to become a vehicle of content generation, sharing and retrieval, demonstrated by such extremely hot applications as P2P live and on-demand streaming, User-Generated Content (UGC) services (e.g., Youtube, Flickr ) and Social Networks (e.g., Facebook, Google+). To match such transformation, a clean-slate redesign of the future Internet architecture has been proposed under the generic name of Content-Centric Networking (CCN). In this paper, we propose a new caching scheme for such CCN networks and evaluate the in-network caching performance of this policy by comparing it with that of the default proposed policy via simulation.

Interest packets retransmission in lossy CCN networks and its impact on network performance

When both interest packets and data chunks can be dropped on the way, due to network impairments, deciding who, the CCN router or the end-system, sets the timeout duration and retransmits the interest packet after a timeout are two key issues that affect the performance of the network. More specifically, they impact directly the occupancy of the pending interest table (PIT). The standard does not address these issues clearly and the typical CCN implementations (like CCNx) address them naively leaving room for further improvements. In lossy networks, if the router does not retransmit pending interests (no-rtx) the average PIT entry lifetime increases dramatically. Conversely, if the CCN router retransmits pending interests (rtx) periodically, it is not clear how frequently it should do so. In this paper we investigate the performance of the two types of routers in lossy networks, in the presence of both caching and interest aggregation. The study aims at shedding some light on how much performance improvement is achieved by one type of routers over the other. We also introduce a new method for estimating the PIT entry timeout that is shown to perform better than the currently used default method in CCN.

Content multi-homing: An alternative approach

The CDN serves as an essential element in providing content delivery services on the Internet; however, limited by its footprint and influenced by variations of network conditions and user demands, no individual CDN can qualitatively fulfill delivery tasks anytime and anywhere. As such, large content providers often use multiple CDNs. Nevertheless, this approach is not only cumbersome to negotiate multiple business contracts, but also economically inefficient, especially for small content providers. A simpler alternative approach would be to let each content provider only sign one contract with an authoritative CDN, say CDN-0, and let CDN-0 handle the potential performance trap of individual CDN. To fulfill its business commitment, CDN-0 could either use its own infrastructure or rent services from other CDNs, which is sometimes indispensable. To make an optimal operating plan for CDN-0 operator, we propose a new problem, called MCDN-CM, whose objective is to minimize CDN-0s operating cost. MCDN-CM is a concave minimization problem and we take advantage of the special form of the practical CDN-pricing function - piece-wise linear concave - to arrive at an optimal solution through an iterative procedure. We conduct numerical experiments under realistic settings and show via the experimental results that CDN-0 can achieve a tremendous cost-saving by using our proposed algorithm.

SDN-based multi-class QoS-guaranteed inter-data center traffic management

When allocating network bandwidth to multiple classes of applications in inter-data center communication, coordination always yields a better utilization of the backbone network. Yet, it often comes at a prohibitively heavy computational and communication cost, making it thus far not a practically viable approach. SDN helped in bridging the communication cost gap by enabling centralized control, and SDN has been recently applied in such inter-DC traffic management. However, the computational cost is still an issue as the efficient and fast response of the centralized traffic engineering algorithm has become crucial to the practicality of such SDN-based approach. In this paper, we present MCTEQ, a utility-optimization-based joint-bandwidth allocation for inter-DC communication with multiple traffic classes, that handles priorities between traffic classes in a soft manner and explicitly considers the delay requirement of Interactive flows. MCTEQ being NP-hard, we apply approximation techniques to lean on the mature and efficient LP solver and obtain fast and accurate approximations. We demonstrate via experiments with Googles inter-DC backbone topology that MCTEQ achieves about 160 Gbps higher network utilization than the existing SWAN solution, yet runs 2.5 times faster. In particular, MCTEQ guarantees that the allocated bandwidth for Interactive flows strictly meets their end-to-end delay requirements.

Optimal resource sharing in multi-cell wireless LANs: Joint channel assignment, association control, airtime sharing and contention resolution

In this paper we study the problem of throughput maximization, with fairness guarantee, in multi-cell WLANs with hidden terminals. In such a scenario, the throughput gained by each flow is generally determined by the four major control mechanisms that make the basis of WLAN control protocols - viz., radio frequency channel assignment, association control, intra-AP airtime sharing between the associated clients, and the access rate control which is determined by the MAC layer contention window tuning. Instead of focusing on each of the four mechanisms apart, in this paper we optimize the network performance via a joint channel assignment, association control, airtime sharing, and contention window tuning; we formulate this joint problem as a non-linear programming problem. Because such a problem is in general NP hard due to the integral nature of both channel assignment and association control, we solve it by iteratively approximating it via a convex programming problem. Numerical and simulation results show that, the resulting solution significantly improves the network throughput and fairness, compared with methods that optimize separately the different mechanisms. Our solution also gives further insights into the importance of each one of these mechanisms when considered jointly.

Energy-efficient planning of QoS-constrained virtual-cluster embedding in data centres

It is crucial for cloud service providers to fulfil their tenants targeted quality of service (QoS) requirements, yet, providing such QoS in an energy efficient manner is of paramount importance to optimizing operational costs. In this paper, we study the energy efficiency of virtual cluster embedding in data centers. More specifically, we focus on designing fast algorithms to place in an energy efficient manner virtual clusters, including virtual machines, virtual switches and virtual topology with bandwidth requirements. We first model the problem as an integer programming problem, and prove its NP hardness, then based on a relaxed version of this model, we propose an approximate algorithm to solve it efficiently.

Iterative greedy algorithms for energy efficient LTE small cell networks

The increasing demand for high bandwidth in cellular wireless networks led to a dramatic increase in the number of deployed LTE femtocell base stations (FBS) in recent years. While the energy consumed by a single FBS is relatively negligible, the collective energy footprint of all FBSs deployed by an operator turns out to be huge. Energy-efficient protocols are needed thus to balance the trade-off between energy saving and bandwidth utilization. The key component of the desired protocol to prevent both energy and bandwidth waste is interference mitigation, which, most previous work has failed to properly consider. To this end, in this paper, we manipulate user equipment (UE) association and OFDMA scheduling with a combination of interference mitigation to enable an energy efficient QoS-constrained medium access for femtocell networks. Recognizing the NP-hardness of the problem, we propose two algorithms with guaranteed convergence. Extensive simulations show that our algorithms outperform alternatives in multiple metrics such as utility, power consumption, and convergence speed.

Regret-based learning for medium access in LTE femtocell networks

With the increased density of LTE femtocell base station (FBSs) deployments, mutual interference between FBSs becomes a bottleneck to network performance. Inter-femtocell coordination has been invoked in the past to address this problem, however it is not practically possible as LTE FBSs do not possess the X2 interface to enable them to communicate with each other. In this paper, we opt for a fully distributed approach and introduce a self-learning MAC protocol for FBSs to allocate resources based on historical user feedback. In our approach, each FBS independently learns about the asymmetric contention level on its links, and transfers resources from highly-contended links to low-contention ones to avoid wasting resources. The learning process is modelled as a cluster-structural regret-based game and simulation results show that our algorithm outperforms alternative algorithms from the literature on a wide range of metrics.

Content peering in content centric networks

The settlement-free peering relationships play a vital role in todays Internet, notably in helping ISPs cope with the dramatic increase in traffic load caused by the recent surge in the demand for videos and user generated content. Because of the added caching capability of CCN, peering in CCN can be expanded to encompass not only content that is permanently stored in the ISPs network, but also content that is temporarily cached at CCN routers in the ISPs network. Intuitively, this content-level peering is likely to benefit the networks involved in the peering relationship; however, this comes at a non-trivial cost. In this paper, we try to answer the question of whether the additional overhead and complexity to extend content-level peering is justifiable in view of the benefits it brings. To this end, we formulate the content-level peering problem as an optimization problem to study its maximum potential benefit. We conduct extensive numerical experiments to evaluate the potential peering benefit under realistic AS-level peering graphs, using realistic video traffic. The experimental results show that the interconnectivity of the peering graph significantly affects the maximum benefit of content-level peering. Compared to local greedy caching, cooperative caching can bring higher peering benefits; yet it is sensitive to parameters like the peering link bandwidth and the AS-level cache size.