Leveraging Zipf's law for traffic offloading

Sarrar, Nadi; Uhlig, Steve; Feldmann, Anja; Sherwood, Rob; Huang, Xin

doi:10.1145/2096149.2096152

Cited by 110 publications

(59 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…First, an algorithm to solve the independent case can be applied to the independent subtrees. Moreover, empirical data shows that while Internet routers typically define a default route (an empty prefix), the prefix hierarchy is typically very flat [15]: prefixes hardly overlap with more than one other prefix. As of February 2013, the Internet-wide BGP routing table contains more than 440k prefixes.…”

Section: The Modelmentioning

confidence: 99%

“…Frequent FIB updates are problematic as upon each update, internal FIB structures have to be rebuilt to ensure routing consistency. In particular, update costs are also critical in the context of Software-Defined Networks (SDN) (e.g., based on OpenFlow [10]), as the network controller is remote from the switch and FIB updates may have to be transmitted over a bandwidth-limited network [15].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Competitive FIB Aggregation for Independent Prefixes: Online Ski Rental on the Trie

Bieńkowski

Schmid

2013

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Abstract. This paper presents an asymptotically optimal online algorithm for compressing the Forwarding Information Base (FIB) of a router under a stream of updates (namely insert rule, delete rule, and change port of prefix). The objective of the algorithm is to dynamically aggregate forwarding rules into a smaller but equivalent set of rules while taking into account FIB update costs. The problem can be regarded as a new variant of ski rental on the FIB trie, and we prove that our deterministic algorithm is 3.603-competitive. Moreover, a lower bound of 1.636 is derived for any online algorithm.

show abstract

Section: The Modelmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Competitive FIB Aggregation for Independent Prefixes: Online Ski Rental on the Trie

Bieńkowski

Schmid

2013

Lecture Notes in Computer Science

View full text Add to dashboard Cite

show abstract

“…Likewise, traffic engineering may benefit from focusing on the few large flows that dominate the bandwidth consumption [92]. Even routing can benefit from offloading the large flows to a fast hardware forwarding device, thereby freeing capacity on the slower software-based router [590].…”

Section: Discussionmentioning

confidence: 99%

Workload Modeling for Computer Systems Performance Evaluation

Feitelson

2015

191

131

View full text Add to dashboard Cite

Reliable performance evaluations require the use of representative workloads. This is no easy task since modern computer systems and their workloads are complex, with many interrelated attributes and complicated structures. Experts often use sophisticated mathematics to analyze and describe workload models, making these models difficult for practitioners to grasp. This book aims to close this gap by emphasizing the intuition and the reasoning behind the definitions and derivations related to the workload models. It provides numerous examples from real production systems, with hundreds of graphs. Using this book, readers will be able to analyze collected workload data and clean it if necessary, derive statistical models that include skewed marginal distributions and correlations, and consider the need for generative models and feedback from the system. The descriptive statistics techniques covered are also useful for other domains.

show abstract

“…FIB update processing delays can lead to limitations in the number of updates per second that can be applied, as prior work has shown is the case for some existing OpenFlow implementations [13]. Our work can also be beneficial in combination with the caching-based IP router design that leverages traffic properties and OpenFlow [16].…”

Section: Fib Aggregation and Software Defined Networkingmentioning

confidence: 92%

Leveraging locality for FIB aggregation

Sarrar

Wuttke

Schmid

et al. 2014

2014 IEEE Global Communications Conference

Self Cite

View full text Add to dashboard Cite

Abstract-Snapshots of the Forwarding Information Base (FIB) in Internet routers can be compressed (or aggregated) to at least half of their original size, as shown by previous studies. However, the permanent stream of updates to the FIB due to routing updates complicates FIB aggregation in practice: keeping a (near-)optimally aggregated FIB in face of these routing updates is algorithmically challenging. A sensible trade-off has to be found between the aggregation gain and the complexity of handling routing updates. This paper investigates whether the spatial and temporal locality properties of routing updates conceal opportunities for improving this trade-off in online FIB aggregation.Our contributions include an empirical study of the locality of updates in public Internet routing data. To facilitate this study, we design the Locality-aware FIB Aggregation (LFA) algorithm. We show, that an algorithm as simple as LFA can effectively leverage the locality of FIB churn to keep low the number of updates to the aggregated FIB, as within time periods of a few seconds or minutes, routing updates affect only a limited number of regions in the FIB. I. INTRODUCTIONAt a high level, Internet routers are built around a route processor which runs routing protocols and makes routing decisions, and a forwarding plane which forwards packets according to the decisions of the route processor. The crucial link between the two components is the Forwarding Information Base (FIB), containing the forwarding rules. The route processor inserts and deletes FIB entries according to its route computations. The forwarding plane uses the FIB to perform an IP destination lookup on each incoming packet. This requires the FIB to support very fast IP destination lookups so that packets can be forwarded at line-rate. In addition, the FIB needs to support frequent updates to the forwarding rules due to the churn in the BGP routing table. Finally, the number of forwarding rules that a FIB needs to keep is growing over time, putting extra pressure on the FIB memory capacity. 1 To fulfill all these requirements, FIB memory used in today's enterprise-grade to high-end routers is a highly specialized component, expensive, and power-hungry. It is also considered a main limiting factor in terms of a router's lifetime [12].A natural and local solution to mitigate the problembefore possible long-term solutions are deployed -is the aggregation (or compression) of the FIB, i.e., the replacement of the existing set of rules by an equivalent but smaller set. The aggregation of FIB rules has the appealing property that it is a purely local solution, in the sense that it does not affect neighboring routers and it can be realized by the route processor through only software modifications.

show abstract

Leveraging Zipf's law for traffic offloading

Cited by 110 publications

References 27 publications

Competitive FIB Aggregation for Independent Prefixes: Online Ski Rental on the Trie

Competitive FIB Aggregation for Independent Prefixes: Online Ski Rental on the Trie

Workload Modeling for Computer Systems Performance Evaluation

Leveraging locality for FIB aggregation

Contact Info

Product

Resources

About