TOFEC: Achieving optimal throughput-delay trade-off of cloud storage using erasure codes

By Guanfeng Liang, U. Kozat. 2013

77 citations recorded by the source corpus.

Our paper presents solutions using erasure coding, parallel connections to storage cloud and limited chunking (i.e., dividing the object into a few smaller segments) together to significantly improve the delay performance of uploading and downloading data in and out of cloud storage. TOFEC is a strategy that helps front-end proxy adapt to level of workload by treating scalable cloud storage (e.g. Amazon S3) as a shared resource requiring admission control. Under light workloads, TOFEC creates more smaller chunks and uses more parallel connections per file, minimizing service delay. Under heavy workloads, TOFEC automatically reduces the level of chunking (fewer chunks with increased size) and uses fewer parallel connections to reduce overhead, resulting in higher throughput and preventing queueing delay. Our trace-driven simulation results show that TOFEC's adaptation mechanism converges to an appropriate code that provides the optimal delay-throughput trade-off without reducing system capacity. Compared to a non-adaptive strategy optimized for throughput, TOFEC delivers 2.5× lower latency under light workloads; compared to a non-adaptive strategy optimized for latency, TOFEC can scale to support over 3× as many requests.

Subfields: Computer Networks and Communications

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Networks and Communications

Citations in the corpus

Speeding Up Distributed Machine Learning Using Codes (2015, methodology): “This model is motivated by the model proposed in [80]: the authors used this distribution to model latency of file queries from cloud storage systems.”
Coded Computation Over Heterogeneous Clusters (2019, methodology): “The shifted exponential model for computation time, which is the sum of a constant (deterministic) term and a variable (stochastic) term, is motivated by the distribution model proposed by Liang and Kozat [33] for latency in querying data files from cloud storage systems.”
On Large-Cohort Training for Federated Learning (2021, methodology): “Such models were found to be good models of runtimes for ﬁle queries in cloud storage systems (Liang and Kozat, 2014) and mini-batch SGD on distributed compute systems (Lee et al., 2017).”
Minimizing latency for secure distributed computing (2017, methodology): “2Our analysis can be generalized to the shifted exponential model used in [6], [13] as we detail in [20].”
Coded computation over heterogeneous clusters (2017, methodology): “The shifted exponential model for computation time, which is the sum of a constant (deterministic) term and a variable (stochastic) term, is motivated by the distribution model p roposed by authors in [11] for latency in querying data files f rom cloud storage systems.”
Improving Distributed Gradient Descent Using Reed-Solomon Codes (2017, background): “Previous works [6], [16] model the return time of a machine as a shifted exponential random variable.”
Secure Distributed Computing With Straggling Servers Using Polynomial Codes (2019, background): “We assume that the computing time distribution Pr(TC ≤ t) follows an exponential distribution [25], i.”