Salmani, Mehran ; Ghafouri, Saeid ; Sanaee, Alireza ; Razavi, Kamran ; Mühlhäuser, Max ; Doyle, Joseph ; Jamshidi, Pooyan ; Sharifi, Mohsen (2023)
Reconciling High Accuracy, Cost-Efficiency, and Low Latency of Inference Serving Systems.
3rd Workshop on Machine Learning and Systems. Rome, Italy (08.05.2023-08.05.2023)
doi: 10.1145/3578356.3592578
Konferenzveröffentlichung, Bibliographie
Kurzbeschreibung (Abstract)
The use of machine learning (ML) inference for various applications is growing drastically. ML inference services engage with users directly, requiring fast and accurate responses. Moreover, these services face dynamic workloads of requests, imposing changes in their computing resources. Failing to right-size computing resources results in either latency service level objectives (SLOs) violations or wasted computing resources. Adapting to dynamic workloads considering all the pillars of accuracy, latency, and resource cost is challenging. In response to these challenges, we propose InfAdapter, which proactively selects a set of ML model variants with their resource allocations to meet latency SLO while maximizing an objective function composed of accuracy and cost. InfAdapter decreases SLO violation and costs up to 65 and 33, respectively, compared to a popular industry autoscaler (Kubernetes Vertical Pod Autoscaler).
| Typ des Eintrags: | Konferenzveröffentlichung |
|---|---|
| Erschienen: | 2023 |
| Autor(en): | Salmani, Mehran ; Ghafouri, Saeid ; Sanaee, Alireza ; Razavi, Kamran ; Mühlhäuser, Max ; Doyle, Joseph ; Jamshidi, Pooyan ; Sharifi, Mohsen |
| Art des Eintrags: | Bibliographie |
| Titel: | Reconciling High Accuracy, Cost-Efficiency, and Low Latency of Inference Serving Systems |
| Sprache: | Deutsch |
| Publikationsjahr: | 8 Mai 2023 |
| Verlag: | ACM |
| Buchtitel: | EuroMLSys '23: Proceedings of the 3rd Workshop on Machine Learning and Systems |
| Veranstaltungstitel: | 3rd Workshop on Machine Learning and Systems |
| Veranstaltungsort: | Rome, Italy |
| Veranstaltungsdatum: | 08.05.2023-08.05.2023 |
| DOI: | 10.1145/3578356.3592578 |
| Kurzbeschreibung (Abstract): | The use of machine learning (ML) inference for various applications is growing drastically. ML inference services engage with users directly, requiring fast and accurate responses. Moreover, these services face dynamic workloads of requests, imposing changes in their computing resources. Failing to right-size computing resources results in either latency service level objectives (SLOs) violations or wasted computing resources. Adapting to dynamic workloads considering all the pillars of accuracy, latency, and resource cost is challenging. In response to these challenges, we propose InfAdapter, which proactively selects a set of ML model variants with their resource allocations to meet latency SLO while maximizing an objective function composed of accuracy and cost. InfAdapter decreases SLO violation and costs up to 65 and 33, respectively, compared to a popular industry autoscaler (Kubernetes Vertical Pod Autoscaler). |
| Freie Schlagworte: | machine learning, inference serving systems, autoscaling |
| Fachbereich(e)/-gebiet(e): | 20 Fachbereich Informatik 20 Fachbereich Informatik > Telekooperation |
| TU-Projekte: | DFG|SFB1053|SFB1053 TPA01 Mühlhä DFG|SFB1053|SFB1053 TPB02 Mühlhä |
| Hinterlegungsdatum: | 02 Aug 2023 14:09 |
| Letzte Änderung: | 04 Aug 2023 07:32 |
| PPN: | 510353878 |
| Export: | |
| Suche nach Titel in: | TUfind oder in Google |
 |
Frage zum Eintrag |
Optionen (nur für Redakteure)
 |
Redaktionelle Details anzeigen |
Drucken
Impressum