Grant: Discovery Grants Program - Individual
The University of British Columbia
$48,000.00
Probabilistic Guarantees for Real-Time Systems
EF 2024-25
Résumé
Climate change, driven by greenhouse gas emissions, is one of today's most important societal challenges. Computing technologies may mitigate some emissions, but the rapid increase in the energy needs of the same systems, and their embodied CO2-equivalent emissions, needs immediate attention. The long-term objective of this proposal is to answer the question: "How can we reach net-zero carbon emissions for our computing platforms?"
Broadly, efforts towards net-zero computing can be mapped to the following avenues for exploration:
1. Improve resource utilization.
2. Improve energy efficiency.
3. Renewable energy matching.
4. Extend device lifetimes.
5. System redesign.
6. Reduce demand.
In this proposal, we choose to focus on the following short term objectives:
- Improving resource utilization. Resource allocation & scheduling decisions are typically made using either worst-case execution time estimates or using average-case execution time estimates. We can make better use of resources by considering the entire execution time distribution of a task. Doing so will allow us to collocate workload better with no impact on timeliness expectations. Because timeliness expectations are often prescribed with percentiles (such as: 99.9% of tasks complete within 250ms), we currently also lack analytical tools to allocate resources effectively.
- Extending device lifetimes. In the overall lifecycle of a device, the manufacturing and end-of-life handling contributes significantly to the embodied carbon costs. We need
to use devices for longer if we have to get closer to net-zero systems. As devices age, they become failure-prone and require additional maintenance. We need better mechanisms for fault-tolerance and graceful service degradation. Proactive failure prediction and maintenance are also crucial.
- Reducing demand. Not all computation needs to be exact. For many applications, end-users tolerate some long-latency operations and can work with approximate results. Using
approximations in place of exact computations can lead to significant demand-side reductions, which reduce the operational carbon emissions (typically through reduced energy consumption). We can try to shape demand, including matching energy availability with the type of computation (e.g., if energy is not available from a renewable source then we would use approximate computation). The work on using approximate computation has synergy with (O1) and (O2). We can resort to approximate computation in the presence of faults to ensure system availability with degraded service. We can also use approximate computation for efficient resource utilization.
The proposed program will involve mathematical analysis, algorithm development (for problems that are usually NP-Hard), simulation studies, as well as system building and evaluation. We will use benchmarks and workload traces (from production systems) to illustrate the gains from our work.
Programme
Discovery Grants Program - Individual
Bénéficiaire
Emplacement
British Columbia, Canada
Type de bénéficiaire
Academia
Type
Grant
Source
NSERC's Awards Data
Tel que publié
Application id
RGPIN-2024-06696
Name nom
Gopalakrishnan, Sathish
Department département
Electrical and Computer Engineering
Organization id
2
Institution établissement
The University of British Columbia
Province en
British Columbia
Province fr
Colombie-Britannique
Country en
CANADA
Country fr
CANADA
Fiscal year exercice financier
2024
Competition year année de concours
2024
Award amount
48000
Program id
RGPIN
Program name en
Discovery Grants Program - Individual
Program name fr
Programme de subventions à la découverte - individuelles
Group en
DISCOVERY RESEARCH
Group fr
RECHERCHE AXÉE SUR LA DÉCOUVERTE
Committee code
1507
Committee name en
Computer Science
Committee name fr
Sciences informatiques
Area of application code
802
Area of application group en
Information and communication services
Area of application group fr
Services d'information et de télécommunications
Area of application en
Information systems and technology
Area of application fr
Systèmes d'information et technologie de l'information
Research subject code
2713
Research subject group en
Information technology
Research subject group fr
Technologie de l'information
Research subject en
Algorithms
Research subject fr
Algorithmes
Field of research list names en
Not available
Field of research list names fr
Non disponible
Socioeconomic objective list names en
Not available
Socioeconomic objective list names fr
Non disponible
Application title
Probabilistic Guarantees for Real-Time Systems
Keywords
real-time systems; scheduling; resource allocation; carbon-aware computing; energy efficiency; net-zero systems
Application summary
Climate change, driven by greenhouse gas emissions, is one of today's most important societal challenges. Computing technologies may mitigate some emissions, but the rapid increase in the energy needs of the same systems, and their embodied CO2-equivalent emissions, needs immediate attention. The long-term objective of this proposal is to answer the question: "How can we reach net-zero carbon emissions for our computing platforms?"
Broadly, efforts towards net-zero computing can be mapped to the following avenues for exploration:
1. Improve resource utilization.
2. Improve energy efficiency.
3. Renewable energy matching.
4. Extend device lifetimes.
5. System redesign.
6. Reduce demand.
In this proposal, we choose to focus on the following short term objectives:
- Improving resource utilization. Resource allocation & scheduling decisions are typically made using either worst-case execution time estimates or using average-case execution time estimates. We can make better use of resources by considering the entire execution time distribution of a task. Doing so will allow us to collocate workload better with no impact on timeliness expectations. Because timeliness expectations are often prescribed with percentiles (such as: 99.9% of tasks complete within 250ms), we currently also lack analytical tools to allocate resources effectively.
- Extending device lifetimes. In the overall lifecycle of a device, the manufacturing and end-of-life handling contributes significantly to the embodied carbon costs. We need
to use devices for longer if we have to get closer to net-zero systems. As devices age, they become failure-prone and require additional maintenance. We need better mechanisms for fault-tolerance and graceful service degradation. Proactive failure prediction and maintenance are also crucial.
- Reducing demand. Not all computation needs to be exact. For many applications, end-users tolerate some long-latency operations and can work with approximate results. Using
approximations in place of exact computations can lead to significant demand-side reductions, which reduce the operational carbon emissions (typically through reduced energy consumption). We can try to shape demand, including matching energy availability with the type of computation (e.g., if energy is not available from a renewable source then we would use approximate computation). The work on using approximate computation has synergy with (O1) and (O2). We can resort to approximate computation in the presence of faults to ensure system availability with degraded service. We can also use approximate computation for efficient resource utilization.
The proposed program will involve mathematical analysis, algorithm development (for problems that are usually NP-Hard), simulation studies, as well as system building and evaluation. We will use benchmarks and workload traces (from production systems) to illustrate the gains from our work.