Sergey Gorinsky
IMDEA
61 Papers
461 Citations
Sergey Gorinsky is an academic researcher from IMDEA. The author has contributed to research in topics: Network congestion & The Internet. The author has an hindex of 18, co-authored 61 publications. Previous affiliations of Sergey Gorinsky include Washington University in St. Louis & University of Texas at Austin.
Chat about Author
Papers
Generalized Multiframe Tasks
TL;DR: A framework for determining feasibility for a wide variety of task systems is established and a framework is applied to this task model to obtain a feasibility-testing algorithm that runs in time pseudo-polynomial in the size of the input for all systems of such tasks whose densities are bounded by a constant less than one.
271
Trade-offs in optimizing the cache deployments of CDNs
Syed Hasan,Sergey Gorinsky,Constantine Dovrolis,Ramesh K. Sitaraman +3 more
- 08 Jul 2014
TL;DR: This paper solves the cache deployment optimization (CaDeOp) problem of determining how much server, energy, and bandwidth resources to provision in each cache AS, i.e., each AS chosen for cache deployment, as a mixed integer program (MIP).
89
RL-Cache: Learning-Based Cache Admission for Content Delivery
Vadim Kirilin,Aditya Sundarrajan,Sergey Gorinsky,Ramesh K. Sitaraman +3 more
- 14 Aug 2019
TL;DR: A novel algorithm called RL-Cache is proposed that uses model-free reinforcement learning (RL) to decide whether or not to admit a requested object into the CDN's cache and weights a large set of features that include the object size, recency, and frequency of access.
79
RL-Cache: Learning-Based Cache Admission for Content Delivery
TL;DR: A novel algorithm called RL-Cache is proposed that uses model-free reinforcement learning (RL) to decide whether or not to admit a requested object into the CDN’s cache and weights a large set of features that include the object size, recency, and frequency of access.
63
Link buffer sizing: a new look at the old problem
Sergey Gorinsky,Anshul Kantawala,Jonathan S. Turner +2 more
- 27 Jun 2005
TL;DR: It is argued that the problem of link buffer sizing needs a new formulation: design a buffer sizing algorithm that accommodates needs of all Internet applications without engaging IP routers in any additional signaling.