Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Deduplicator: When Computation Reuse Meets Load Balancing at the Network Edge

Md Washik Al Azad, Spyridon Mastorakis

TL;DR

This paper introduces Deduplicator, a middlebox for edge computing that harmonizes computation reuse with load balancing by using Locality Sensitive Hashing to detect similar tasks and adaptively slice the hash-value space across edge servers. It details a design that collects resource usage, handles dynamic server membership, and routes tasks to maximize reuse without creating severe imbalances. Empirical results with real-world datasets show up to 20% higher reuse versus non-reuse-aware baselines, while maintaining line-rate performance and load balance (less than 5% deviation), and reuse accuracy above 90%. The work demonstrates the practical viability of joint reuse-aware routing and load balancing for scalable edge deployments and suggests that adaptive hash-space partitioning can achieve favorable trade-offs between throughput, latency, and computation reuse.

Abstract

Load balancing has been a fundamental building block of cloud and, more recently, edge computing environments. At the same time, in edge computing environments, prior research has highlighted that applications operate on similar (correlated) data. Based on this observation, prior research has advocated for the direction of "computation reuse", where the results of previously executed computational tasks are stored at the edge and are reused (if possible) to satisfy incoming tasks with similar input data, instead of executing incoming tasks from scratch. Both load balancing and computation reuse are critical to the deployment of scalable edge computing environments, yet they are contradictory in nature. In this paper, we propose the Deduplicator, a middlebox that aims to facilitate both load balancing and computation reuse at the edge. The Deduplicator features mechanisms to identify and deduplicate similar tasks offloaded by user devices, collect information about the usage of edge servers' resources, manage the addition of new edge servers and the failures of existing edge servers, and ultimately balance the load imposed on edge servers. Our evaluation results demonstrate that the Deduplicator achieves up to 20% higher percentages of computation reuse compared to several other load balancing approaches, while also effectively balancing the distribution of tasks among edge servers at line rate.

Deduplicator: When Computation Reuse Meets Load Balancing at the Network Edge

TL;DR

This paper introduces Deduplicator, a middlebox for edge computing that harmonizes computation reuse with load balancing by using Locality Sensitive Hashing to detect similar tasks and adaptively slice the hash-value space across edge servers. It details a design that collects resource usage, handles dynamic server membership, and routes tasks to maximize reuse without creating severe imbalances. Empirical results with real-world datasets show up to 20% higher reuse versus non-reuse-aware baselines, while maintaining line-rate performance and load balance (less than 5% deviation), and reuse accuracy above 90%. The work demonstrates the practical viability of joint reuse-aware routing and load balancing for scalable edge deployments and suggests that adaptive hash-space partitioning can achieve favorable trade-offs between throughput, latency, and computation reuse.

Abstract

Load balancing has been a fundamental building block of cloud and, more recently, edge computing environments. At the same time, in edge computing environments, prior research has highlighted that applications operate on similar (correlated) data. Based on this observation, prior research has advocated for the direction of "computation reuse", where the results of previously executed computational tasks are stored at the edge and are reused (if possible) to satisfy incoming tasks with similar input data, instead of executing incoming tasks from scratch. Both load balancing and computation reuse are critical to the deployment of scalable edge computing environments, yet they are contradictory in nature. In this paper, we propose the Deduplicator, a middlebox that aims to facilitate both load balancing and computation reuse at the edge. The Deduplicator features mechanisms to identify and deduplicate similar tasks offloaded by user devices, collect information about the usage of edge servers' resources, manage the addition of new edge servers and the failures of existing edge servers, and ultimately balance the load imposed on edge servers. Our evaluation results demonstrate that the Deduplicator achieves up to 20% higher percentages of computation reuse compared to several other load balancing approaches, while also effectively balancing the distribution of tasks among edge servers at line rate.
Paper Structure (18 sections, 2 equations, 11 figures, 3 tables)

This paper contains 18 sections, 2 equations, 11 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Use Cases and Motivation
  3. Computation Reuse Use Cases
  4. Deduplicator Motivation
  5. Deduplicator Design
  6. Design Overview
  7. Task Offloading and Execution Process
  8. Hash Value Space Slicing
  9. Collecting Resource Usage Information From Edge Servers
  10. Edge Server Additions and Failures
  11. Evaluation
  12. Deduplicator Prototype Implementation
  13. Evaluation Setup
  14. Evaluation Results
  15. Deduplicator Design Tradeoffs
  16. ...and 3 more sections

Figures (11)

  • Figure 1: Similarity among video frames in different application scenarios.
  • Figure 2: An edge computing environment where the Deduplicator receives tasks offloaded by user devices and distributes them to edge servers.
  • Figure 3: Distribution (slicing) of a hash value space (hash length equal to 2 bytes) among three edge servers.
  • Figure 4: Overview of the Deduplicator implementation.
  • Figure 5: Experimental topology.
  • ...and 6 more figures