All welcome and open to the public!
Time: Tuesday, September 25, 2018, 2:30–3:30pm
Location: ECS660
Title: Learning-based Adaptive Data Placement for Low Latency in Data
Center Networks
Speaker: Kaiyang Liu
Abstract: Low-latency data access is an important challenge for data
center networks. Proper placement of the data items can reduce the data
travel time in the distributed storage systems, which contributes
significantly to the latency reduction. Most existing data placement
approaches have often assumed the prior distribution of data requests or
discovered so through trace analysis. However, the traditional static
model-based solutions are less effective to handle the system
uncertainties in a dynamic environment. We present DataBot, a
reinforcement learning-based adaptive framework, to learn the optimal data
placement policies faced with the dynamic network conditions and
time-varying request patterns. DataBot utilizes a neural network, trained
with a variant of Q-learning, whose input is the real-time data flow
measurements and whose output is a value function estimating the
near-future latency. For rapid decision making, DataBot is divided into
two decoupled production and training components, ensuring that the
convergence time of the training will not introduce more overheads to
serve the read/write requests. Evaluation results demonstrate that the
average write and read latency of the whole system can be lowered by about
35% and 40%, respectively.
—
Title: Intelligent Caching in Dense Small-Cell Networks with Limited
External Resources
Speaker: Bingshan Hu
Abstract: A promising solution to alleviate the mobile traffic burden on
the Internet is to cache the most popular content at the heterogeneous
wireless network edge. However, due to the vast content stored at the
remote server, and to cache effectively, it concerns the file popularity
profile that may not be known by the network operators in advance.
Therefore, online learning techniques are used to tackle the challenges
brought by the unknown knowledge. We present an effective and efficient
algorithm based on the stochastic combinatorial multi-armed bandits with
locked-up slots to address the content caching problem. Our work
particularly addresses the scenario where dense small cells with diverse
user populations are deployed. Additionally, this network is only given
limited external resources such as computational resource to learn the
caching policies and wireless back haul resource to refresh the caches.
Our algorithm learns the caching policies online which is to decide which
files to be cached sequentially. Despite sharing the limited external
resources, the proposed algorithm guarantees the performance of each small
cell to approach the optimum. Experiments are conducted to cross-validate
the theorem presented in this work.