Mode of access: World Wide Web.

System requirements: Adobe Acrobat Reader.

Part of: Synthesis digital library of engineering and computer science.

Series from website.

Includes bibliographical references (p. 123-135).

1. Introduction -- 1.1 Warehouse-scale computers -- 1.2 Cost efficiency at scale -- 1.3 Not just a collection of servers -- 1.4 One datacenter vs. several datacenters -- 1.5 Why WSCS might matter to you -- 1.6 Architectural overview of WSCS -- 1.6.1 Storage -- 1.6.2 Networking fabric -- 1.6.3 Storage hierarchy -- 1.6.4 Quantifying latency, bandwidth, and capacity -- 1.6.5 Power usage -- 1.6.6 Handling failures --

2. Workloads and software infrastructure -- 2.1 Datacenter vs. desktop -- 2.2 Performance and availability toolbox -- 2.3 Platform-level software -- 2.4 Cluster-level infrastructure software -- 2.4.1 Resource management -- 2.4.2 Hardware abstraction and other basic services -- 2.4.3 Deployment and maintenance -- 2.4.4 Programming frameworks -- 2.5 Application-level software -- 2.5.1 Workload examples -- 2.5.2 Online: web search -- 2.5.3 Offline: scholar article similarity -- 2.6 A monitoring infrastructure -- 2.6.1 Service-level dashboards -- 2.6.2 Performance debugging tools -- 2.6.3 Platform-level health monitoring -- 2.7 Buy vs. build -- 2.8 Tail-tolerance -- 2.9 Further reading --

3. Hardware Building Blocks -- 3.1 Cost-efficient server hardware -- 3.1.1 The impact of large SMP communication efficiency -- 3.1.2 Brawny vs. wimpy servers -- 3.1.3 Balanced designs -- 3.2 WSC storage -- 3.2.1 Unstructured WSC storage -- 3.2.2 Structured WSC storage -- 3.2.3 Interplay of storage and networking technology -- 3.3 WSC networking -- 3.4 Further reading --

4. Datacenter basics -- 4.1 Datacenter tier classifications and specifications -- 4.2 Datacenter power systems -- 4.2.1 Uninterruptible power systems -- 4.2.2 Power distribution units -- 4.2.3 Alternative: DC distribution -- 4.3 Datacenter cooling systems -- 4.3.1 CRACs, chillers, and cooling towers -- 4.3.2 CRACs -- 4.3.3 Chillers -- 4.3.4 Cooling towers -- 4.3.5 Free cooling -- 4.3.6 Air flow considerations -- 4.3.7 In-rack, in-row cooling, and cold plates -- 4.3.8 Case study: Google's in-row cooling -- 4.3.9 Container-based datacenters -- 4.4 Summary --

5. Energy and power efficiency -- 5.1 Datacenter energy efficiency -- 5.1.1 The PUE metric -- 5.1.2 Issues with the PUE metric -- 5.1.3 Sources of efficiency losses in datacenters -- 5.1.4 Improving the energy efficiency of datacenters -- 5.1.5 Beyond the facility -- 5.2 The energy efficiency of computing -- 5.2.1 Measuring energy efficiency -- 5.2.2 Server energy efficiency -- 5.2.3 Usage profile of warehouse-scale computers -- 5.3 Energy-proportional computing -- 5.3.1 Causes of poor energy proportionality -- 5.3.2 Improving energy proportionality -- 5.3.3 Energy proportionality, the rest of the system -- 5.4 Relative effectiveness of low-power modes -- 5.5 The role of software in energy proportionality -- 5.6 Datacenter power provisioning -- 5.6.1 Deploying the right amount of equipment -- 5.6.2 Oversubscribing facility power -- 5.7 Trends in server energy usage -- 5.7.1 Using energy storage for power management -- 5.8 Conclusions -- 5.8.1 Further reading --

6. Modeling costs -- 6.1 Capital costs -- 6.2 Operational costs -- 6.3 Case studies -- 6.3.1 Real-world datacenter costs -- 6.3.2 Modeling a partially filled datacenter -- 6.3.3 The cost of public clouds --

7. Dealing with failures and repairs -- 7.1 Implications of software-based fault tolerance -- 7.2 Categorizing faults -- 7.3 Machine-level failures -- 7.4 Repairs -- 7.5 Tolerating faults, not hiding them --

8. Closing remarks -- 8.1 Hardware -- 8.2 Software -- 8.3 Economics -- 8.4 Key challenges -- 8.4.1 Rapidly changing workloads -- 8.4.2 Building responsive large scale systems -- 8.4.3 Energy proportionality of non-CPU components -- 8.4.4 Overcoming the end of Dennard scaling -- 8.4.5 Amdahl's cruel law -- 8.5 Conclusions --

Bibliography -- Author biographies.

Abstract freely available; full-text restricted to subscribers or individual document purchasers.

Compendex

INSPEC

Google scholar

Google book search

As computation continues to move into the cloud, the computing platform of interest no longer resembles a pizza box or a refrigerator, but a warehouse full of computers. These new large datacenters are quite different from traditional hosting facilities of earlier times and cannot be viewed simply as a collection of co-located servers. Large portions of the hardware and software resources in these facilities must work in concert to efficiently deliver good levels of Internet service performance, something that can only be achieved by a holistic approach to their design and deployment. In other words, we must treat the datacenter itself as one massive warehouse-scale computer (WSC). We describe the architecture of WSCs, the main factors influencing their design, operation, and cost structure, and the characteristics of their software base. We hope it will be useful to architects and programmers of today's WSCs, as well as those of future many-core platforms which may one day implement the equivalent of today's WSCs on a single board.

Also available in print.

Title from PDF t.p. (viewed on August 14, 2013).