Attila Korosi

Compressing IP Forwarding Tables: Realizing Information-Theoretical Space Bounds and Fast Lookups Simultaneously

The Internet routing ecosystem is facing compelling scalability challenges, manifested primarily in the rapid growth of IP packet forwarding tables. The forwarding table, implemented at the data plane fast path of Internet routers to drive the packet forwarding process, currently contains about half a million entries and counting. Meanwhile, it needs to support millions of complex queries and updates per second. In this paper, we make the curious observation that the entropy of IP forwarding tables is very small and, what is more, seems to increase at a lower pace than the size of the network. This suggests that a sophisticated compression scheme may effectively and persistently reduce the memory footprint of IP forwarding tables, shielding operators from scalability matters at least temporarily. Our main contribution is such a compression scheme which, for the first time, admits both the required information-theoretical size bounds and attains fast lookups, thanks to aggressive level compression. Although we find the underlying optimization problem NP-complete, we can still give a lightweight heuristic algorithm with firm approximation guarantees. This allows us to squeeze real IP forwarding tables, comprising almost 500, 000 prefixes, to just about 140-200 KBytes of memory within a factor of 2-3 of the entropy bound, so that forwarding decisions take only 8-10 memory accesses on average and updates are supported efficiently. Our compression scheme may be of more general interest, as it is applicable to essentially any prefix tree.

Compressing IP forwarding tables for fun and profit

About what is the smallest size we can compress an IP Forwarding Information Base (FIB) down to, while still guaranteeing fast lookup? Is there some notion of FIB entropy that could serve as a compressibility metric? As an initial step in answering these questions, we present a FIB data structure, called Multibit Burrows-Wheeler transform (MBW), that is fundamentally pointerless, can be built in linear time, guarantees theoretically optimal longest prefix match, and compresses to higher-order entropy. Measurements on a Linux prototype provide a first glimpse of the applicability of MBW.

On Distributed Dynamic Spectrum Allocation for Sequential Arrivals

We study the potential of a dynamic spectrum management framework that enables sequential allocation of frequency bands for wireless service providers. We present our distributed system design on allocation and pricing with the goal of achieving efficient spectrum utilization, flexible allocations and incentive-compatibility, considering physical interference model among frequency licensees. Our work provides insights on the emerging optimization problems related to the allocation. We show the analytic results that give ideas for well-established heuristics to these problems. We evaluate the proposed framework and algorithms numerically, and we arrive at the conclusion, that even by the simple proposed heuristics the system performs well, thus it is a suitable approach to a flexible distributed dynamic allocation framework.

On greedy network formation

Greedy navigability is a central issue in the theory of networks. However, the exogenous nature of network models do not allow for describing how greedy routable-networks emerge in reality. In turn, network formation games focus on the very emergence proess, but the applied shortest-path based cost functions exclude navigational aspects. This paper takes a frst step towards incorporating both emergence (missing in algorithmic network models) and greedy navigability (missing in network formation games) into a single framework, and proposes the Greedy Network Formation Game. Our first contribution is the game definition, where we assume a hidden metric space underneath the network, and, instead of usual shortest path metric, we use the length of greedy paths as the measure of communiation cost between players. Our main finding is that greedy-routable small worlds do not emerge on constant dimensional Eulidean grids. This simply means that the emergence of topologies on which w eunderstood the priniples of greedy forwarding cannot be explained endogenously. We also present a very brief outlook on how the situation hanges in the hyperbolic space.

Analysis on Staleness-Aware Mechanisms of Real-Time Transport Protocols in WPANs

Supporting real-time data transport in highly dynamic conditions, which arise in wireless and mobile ad hoc networks is challenging. The significant variations in link bandwidth and latency undermine the quality of real-time applications by introducing congestion and delay jitters. For such applications data is only useful before its lifetime expires. In this paper we investigate a staleness-aware mechanism for realtime multimedia flows, where the system can decide if the respective frame is eligible or not for transmission, namely, if it can be entirely transmitted before its lifetime expires. A mathematical model and analysis is presented, where the partial and complete frame discarding staleness-aware techniques are compared. Then the Burst Transfer Eligibility Decision (BTED) mechanism, a concrete staleness-aware solution for WPANs is introduced. In order to have a deeper understanding of staleness-aware mechanisms, simulation analysis is also presented. We show that complete frame discarding presents better results than the partial rejection in terms of several performance metrics.

Stochastic Model of Finite Buffer Priority Queuing System with Multi-type Batch Arrival and General Rejection

This paper is motivated by the need of buffer dimensioning and traffic management solution for DSL access lines providing quality assured service delivery. To provide exact performance evaluation of such systems often queuing models are applied. The quality of service analysis of access lines using pre-emption option needs analysis of finite buffer priority queuing systems with batch arrival. The presented work extends this queuing model by introducing general rejection and multi-type arrival. To derive numerical results for performance evaluation of queuing systems based on this model analytical framework is given and general application of the model is proven by showing examples of partial, complete and generally linear rejections.