Awesome and Dark Tactics
Homepage Catalog Tag Selection Contributions
All Tags AWS algorithm-design architecture cloud-principles cost-reduction data-centric data-compression data-processing deployment design edge-computing energy-footprint hardware libraries locality machine-learning management measured migration model-optimization model-training performance queries rebuilding scaling services strategies template workloads

<- Back to category

Tactic: Perform specialized tasks that occur infrequently in the cloud

Tactic sort: Awesome Tactic
Type: Architectural Tactic
Category: resource-allocation
Tags: cloud-principles  migration 

Title

Perform specialized tasks that occur infrequently in the cloud

Description

Specialized tasks that occur infrequently might need specialized hardware. To benefit from economies of scale, it is more efficient to share these resources among more consumers. Training a ML model is an example of a time-consuming but infrequent task that requires many (specific) resources. Hence, it can be more efficient to perform this task in the cloud as cloud consumers do not need to purchase these specific resources for a workload that is not frequently used. A consumer who purchases hardware that is infrequently used and otherwise runs idle has a negative effect on energy efficiency. If multiple consumers share the hardware in the cloud, the hardware will be more efficiently used and, therefore, is expected to have a positive effect on the energy efficiency.

Participant

< unknown >

Related software artifact

Specialized tasks

Context

Migration from on-premise to cloud

Software feature

< unknown >

Tactic intent

To benefit from economies of scale in cloud, to reduce costs of purchasing specialized hardware

Target quality attribute

Cost-efficiency

Other related quality attributes

Energy-efficiency

Measured impact

< unknown >

Source

Master Thesis “Architectural Tactics to Optimize Software for Energy Efficiency in the Public Cloud” by Sophie Vos


Graphical representation

  • Contact person
  • Patricia Lago (VU Amsterdam)
  •  disc at vu.nl
  •  patricialago.nl

The Archive of Awesome and Dark Tactics (AADT) is an initiative of the Digital Sustainability Center (DiSC). It received funding from the VU Amsterdam Sustainability Institute, and is maintained by the S2 Group of the Vrije Universiteit Amsterdam.

Initial development of the Archive of Awesome and Dark Tactics by Robin van der Wiel