Learning to Cache and Caching to Learn: Regret Analysis of Caching Algorithms

Crucial performance metrics of a caching algorithm include its ability to quickly and accurately learn a popularity distribution of requests. However, a majority of work on analytical performance analysis focuses on hit probability after an asymptotically large time has elapsed. We consider an online learning viewpoint, and characterize the "regret" in terms of the finite time difference between the hits achieved by a candidate caching algorithm with respect to a genie-aided scheme that places the most popular items in the cache. We first consider the Full Observation regime wherein all requests are seen by the cache. We show that the Least Frequently Used (LFU) algorithm is able to achieve order optimal regret, which is matched by an efficient counting algorithm design that we call LFU-Lite. We then consider the Partial Observation regime wherein only requests for items currently cached are seen by the cache, making it similar to an online learning problem related to the multi-armed bandit problem. We show how approaching this "caching bandit" using traditional approaches yields either high complexity or regret, but a simple algorithm design that exploits the structure of the distribution can ensure order optimal regret. We conclude by illustrating our insights using numerical simulations.

[1]  J. Kiefer,et al.  Asymptotic Minimax Character of the Sample Distribution Function and of the Classical Multinomial Estimator , 1956 .

[2]  Shipra Agrawal,et al.  Further Optimal Regret Bounds for Thompson Sampling , 2012, AISTATS.

[3]  Deniz Gündüz,et al.  Learning-based optimization of cache content in a small cell base station , 2014, 2014 IEEE International Conference on Communications (ICC).

[4]  Peter Auer,et al.  Finite-time Analysis of the Multiarmed Bandit Problem , 2002, Machine Learning.

[5]  Stratis Ioannidis,et al.  Jointly Optimal Routing and Caching for Arbitrary Network Topologies , 2017, IEEE Journal on Selected Areas in Communications.

[6]  Michele Garetto,et al.  A unified approach to the performance analysis of caching systems , 2013, IEEE INFOCOM 2014 - IEEE Conference on Computer Communications.

[7]  W. R. Thompson ON THE LIKELIHOOD THAT ONE UNKNOWN PROBABILITY EXCEEDS ANOTHER IN VIEW OF THE EVIDENCE OF TWO SAMPLES , 1933 .

[8]  Petros Zerfos,et al.  Platform and applications for massive-scale streaming network analytics , 2013, IBM J. Res. Dev..

[9]  Nimrod Megiddo,et al.  ARC: A Self-Tuning, Low Overhead Replacement Cache , 2003, FAST.

[10]  Benny Van Houdt,et al.  Asymptotically Exact TTL-Approximations of the Cache Replacement Algorithms LRU(m) and h-LRU , 2016, 2016 28th International Teletraffic Congress (ITC 28).

[11]  Florin Ciucu,et al.  Exact analysis of TTL cache networks , 2014, Perform. Evaluation.

[12]  Hiroshi Nakagawa,et al.  Optimal Regret Analysis of Thompson Sampling in Stochastic Multi-armed Bandit Problem with Multiple Plays , 2015, ICML.

[13]  Erol Gelenbe,et al.  A Unified Approach to the Evaluation of a Class of Replacement Algorithms , 1973, IEEE Transactions on Computers.

[14]  David Tse,et al.  Probabilistic methods for web caching , 2001, Perform. Evaluation.

[15]  Shipra Agrawal,et al.  Thompson Sampling for Contextual Bandits with Linear Payoffs , 2012, ICML.

[16]  Roy Friedman,et al.  TinyLFU: A Highly Efficient Cache Admission Policy , 2014, PDP.

[17]  Sébastien Bubeck,et al.  Regret Analysis of Stochastic and Nonstochastic Multi-armed Bandit Problems , 2012, Found. Trends Mach. Learn..

[18]  Jiangchuan Liu,et al.  Statistics and Social Network of YouTube Videos , 2008, 2008 16th Interntional Workshop on Quality of Service.

[19]  Yu Gu,et al.  Watch global, cache local: YouTube network traffic at a campus network: measurements and implications , 2008, Electronic Imaging.

[20]  Nicolò Cesa-Bianchi,et al.  Combinatorial Bandits , 2012, COLT.

[21]  T. L. Lai Andherbertrobbins Asymptotically Efficient Adaptive Allocation Rules , 2022 .

[22]  Javad Ghaderi,et al.  Adaptive TTL-Based Caching for Content Delivery , 2017, SIGMETRICS.

[23]  Vianney Perchet,et al.  Bounded regret in stochastic multi-armed bandits , 2013, COLT.

[24]  Ronald Fagin,et al.  Asymptotic Miss Ratios over Independent References , 1977, J. Comput. Syst. Sci..

[25]  W. Hoeffding Probability Inequalities for sums of Bounded Random Variables , 1963 .

[26]  George Karakostas,et al.  Exploitation of different types of locality for Web caches , 2002, Proceedings ISCC 2002 Seventh International Symposium on Computers and Communications.

[27]  Jian Li,et al.  Accurate Learning or Fast Mixing? Dynamic Adaptability of Caching Algorithms , 2017, IEEE Journal on Selected Areas in Communications.

[28]  Peter J. Denning,et al.  Operating Systems Theory , 1973 .

[29]  Hao Che,et al.  Hierarchical Web caching systems: modeling, design and experimental results , 2002, IEEE J. Sel. Areas Commun..

[30]  Donald F. Towsley,et al.  Approximate Models for General Cache Networks , 2010, 2010 Proceedings IEEE INFOCOM.