Overview of WSC Workloads
Open Access
- 28 December 2025
- book chapter
- Published by Springer Nature
- p. 19-53
Abstract
The applications that run on warehouse-scale computers (WSCs) define the system design tradeoffs that we will be discussing in the rest of this book. Hence, we first discuss these applications in this chapter. We start with an overview of broad workload classes including both hyperscale and cloud compute consumption at Google. We then discuss the design of some key workloads: Search, YouTube, Gmail, and key software platforms for machine learning and data processing, and Google Cloud Platform. From these examples, we highlight the key distinguishing characteristics of software that run in WSCs and identify common themes specific to large-scale distributed systems, and discuss key commonalities and differences between hyperscale and cloud workloads.Keywords
This publication has 19 references indexed in Scilit:
- Profiling a warehouse-scale computer2015
- Large-scale cluster management at Google with Borg2015
- The tail at scaleCommunications of the ACM, 2013
- FlumeJava2010
- Eventually ConsistentQueue, 2008
- MapReduceCommunications of the ACM, 2008
- Dynamo2007
- Paxos made live2007
- Lessons from giant-scale servicesIEEE Internet Computing, 2001
- The anatomy of a large-scale hypertextual Web search engineComputer Networks and ISDN Systems, 1998