Definição de Acordos de Nível de Serviço Personalizados para Treinamento de Modelos via Consultas k-Skyband
Resumo
O treinamento de modelos de Aprendizado de Máquina (ML) envolve a configuração de parâmetros, hiperparâmetros e do ambiente de execução, o que resulta em tempos de execução e consumo de energia distintos. Em ambientes de nuvem, os usuários precisam definir Acordos de Nível de Serviço (SLA), escolhendo recursos computacionais com base nas características do treinamento. Neste artigo, estendemos o conceito de Personalized Service Level Agreement (PSLA), propondo o PSLA4ML, que permite especificar requisitos e obter configurações de recursos, com custos e métricas associados ao workflow de treinamento do modelo. Os PSLA4ML são gerados por meio de consultas 𝑘-Skyband, considerando múltiplas métricas, como tempo, custo financeiro , etc. A abordagem foi avaliada por meio de workflows de treinamento de modelos da família BERT, o que indica sua viabilidade.
Palavras-chave:
Acordo de Nível de Serviço Personalizado, Treinamento de Aprendizado de Máquina, Consultas Skyband, Computação em Nuvem, Otimização Multiobjetivo
Referências
Amorim, A., Moraes, J. V., Pina, D., Paes, A., and de Oliveira, D. (2026). On provenance in model fusion. In SBBD 2026 - Short Papers.
Bedo, M., Ciaccia, P., Martinenghi, D., and De Oliveira, D. (2019). A k–skyband approach for feature selection. In SISAP’19, pages 160–168. Springer.
Borzsony, S., Kossmann, D., and Stocker, K. (2001). The skyline operator. In Proceedings 17th international conference on data engineering, pages 421–430. IEEE.
Boukadi, K., Grati, R., and Ben-Abdallah, H. (2016). Toward the automation of a qos-driven sla establishment in the cloud. Service Oriented Computing and Applications, 10(3):279–302.
Ciaccia, P. and Martinenghi, D. (2020). Flexible skylines: Dominance for arbitrary sets of monotone functions. ACM Transactions on Database Systems (TODS), 45(4):1–45.
Ciaccia, P. and Martinenghi, D. (2025). Optimization strategies for parallel computation of skylines. Distributed and Parallel Databases, 43(1):10.
Devlin, J., Chang, M.-W., Lee, K., and Toutanova, K. (2019). Bert: Pre-training of deep bidirectional transformers for language understanding. In Proc. of the 2019 NAACL-HLT, pages 4171–4186.
Gao, Y., Miao, X., Cui, H., Chen, G., and Li, Q. (2014). Processing k-skyband, constrained skyline, and group-by skyline queries on incomplete data. Expert Systems with Applications, 41(10):4959–4974.
Gouryraj, S., Kataria, S., and Swvigaradoss, J. (2021). Service level agreement breach prediction in servicenow. In 2021 Third International Conference on Inventive Research in Computing Applications (ICIRCA), pages 689–698.
Halperin, D., de Almeida, V. T., Choo, L. L., et al. (2014). Demonstration of the myria big data management service. In Dyreson, C. E., Li, F., and Özsu, M. T., editors, SIGMOD’14, pages 881–884. ACM.
Herschel, M., Diestelkämper, R., and Ben Lahmar, H. (2017). A survey on provenance: What for? what form? what from? VLDB J., 26(6):881–906.
Krizhevsky, A., Sutskever, I., and Hinton, G. E. (2017). Imagenet classification with deep convolutional neural networks. Commun. ACM, 60(6):84–90.
Lu, J., Gong, P., Ye, J., Zhang, J., and Zhang, C. (2023). A survey on machine learning from few samples. Pattern Recognition, 139:109480.
Luccioni, A. S., Viguier, S., and Ligozat, A.-L. (2023). Estimating the carbon footprint of bloom, a 176b parameter language model. J. Mach. Learn. Res., 24(1).
Nakandala, S., Zhang, Y., and Kumar, A. (2020). Cerebro: A data system for optimized deep learning model selection. Proceedings of the VLDB Endowment, 13(12).
Ortiz, J., de Almeida, V. T., and Balazinska, M. (2015). Changing the face of database cloud services with personalized service level agreements. In CIDR’15. www.cidrdb.org.
Papadias, D., Tao, Y., Fu, G., and Seeger, B. (2005). Progressive skyline computation in database systems. ACM Trans. Database Syst., 30(1):41–82.
Peoples, C., Moore, A., and Georgalas, N. (2022). Customer classification recommender to support personalised service level agreements across the internet of things. In 2022 IEEE 8th World Forum on Internet of Things (WF-IoT), pages 1–6. IEEE.
Pina, D., Chapman, A., Kunstmann, L., de Oliveira, D., and Mattoso, M. (2024). Dlprov: A data-centric support for deep learning workflow analyses. In Proc. of the DEEM@SIGMOD 2024, page 77–85, New York, NY, USA. ACM.
Pina, D. B., Kunstmann, L. N. O., Chapman, A., de Oliveira, D., and Mattoso, M. (2025). Dlprov: a suite of provenance services for deep learning workflow analyses. PeerJ Comput. Sci., 11:e2985.
Siddique, M. A., Zaman, A., and Morimoto, Y. (2018). Selection of important sets by using k-skyband query for sets. International Journal of Advanced Computer Science and Applications, 9(4).
Talaei Khoei, T., Ould Slimane, H., and Kaabouch, N. (2023). Deep learning: systematic review, models, challenges, and research directions. Neural Computing and Applications, 35(31):23103–23124.
Venkataraman, S., Yang, Z., Franklin, M., Recht, B., and Stoica, I. (2016). Ernest: efficient performance prediction for large-scale advanced analytics. In Proceedings of the 13th Usenix Conference on Networked Systems Design and Implementation, NSDI’16, page 363–378, USA. USENIX Association.
Yao, S., Chen, H., Hanjie, A. W., Yang, R., and Narasimhan, K. (2023). Collie: Systematic construction of constrained text generation tasks.
Zhao, W. X. et al. (2023). A survey of large language models. arXiv:2303.18223, 1(2).
Bedo, M., Ciaccia, P., Martinenghi, D., and De Oliveira, D. (2019). A k–skyband approach for feature selection. In SISAP’19, pages 160–168. Springer.
Borzsony, S., Kossmann, D., and Stocker, K. (2001). The skyline operator. In Proceedings 17th international conference on data engineering, pages 421–430. IEEE.
Boukadi, K., Grati, R., and Ben-Abdallah, H. (2016). Toward the automation of a qos-driven sla establishment in the cloud. Service Oriented Computing and Applications, 10(3):279–302.
Ciaccia, P. and Martinenghi, D. (2020). Flexible skylines: Dominance for arbitrary sets of monotone functions. ACM Transactions on Database Systems (TODS), 45(4):1–45.
Ciaccia, P. and Martinenghi, D. (2025). Optimization strategies for parallel computation of skylines. Distributed and Parallel Databases, 43(1):10.
Devlin, J., Chang, M.-W., Lee, K., and Toutanova, K. (2019). Bert: Pre-training of deep bidirectional transformers for language understanding. In Proc. of the 2019 NAACL-HLT, pages 4171–4186.
Gao, Y., Miao, X., Cui, H., Chen, G., and Li, Q. (2014). Processing k-skyband, constrained skyline, and group-by skyline queries on incomplete data. Expert Systems with Applications, 41(10):4959–4974.
Gouryraj, S., Kataria, S., and Swvigaradoss, J. (2021). Service level agreement breach prediction in servicenow. In 2021 Third International Conference on Inventive Research in Computing Applications (ICIRCA), pages 689–698.
Halperin, D., de Almeida, V. T., Choo, L. L., et al. (2014). Demonstration of the myria big data management service. In Dyreson, C. E., Li, F., and Özsu, M. T., editors, SIGMOD’14, pages 881–884. ACM.
Herschel, M., Diestelkämper, R., and Ben Lahmar, H. (2017). A survey on provenance: What for? what form? what from? VLDB J., 26(6):881–906.
Krizhevsky, A., Sutskever, I., and Hinton, G. E. (2017). Imagenet classification with deep convolutional neural networks. Commun. ACM, 60(6):84–90.
Lu, J., Gong, P., Ye, J., Zhang, J., and Zhang, C. (2023). A survey on machine learning from few samples. Pattern Recognition, 139:109480.
Luccioni, A. S., Viguier, S., and Ligozat, A.-L. (2023). Estimating the carbon footprint of bloom, a 176b parameter language model. J. Mach. Learn. Res., 24(1).
Nakandala, S., Zhang, Y., and Kumar, A. (2020). Cerebro: A data system for optimized deep learning model selection. Proceedings of the VLDB Endowment, 13(12).
Ortiz, J., de Almeida, V. T., and Balazinska, M. (2015). Changing the face of database cloud services with personalized service level agreements. In CIDR’15. www.cidrdb.org.
Papadias, D., Tao, Y., Fu, G., and Seeger, B. (2005). Progressive skyline computation in database systems. ACM Trans. Database Syst., 30(1):41–82.
Peoples, C., Moore, A., and Georgalas, N. (2022). Customer classification recommender to support personalised service level agreements across the internet of things. In 2022 IEEE 8th World Forum on Internet of Things (WF-IoT), pages 1–6. IEEE.
Pina, D., Chapman, A., Kunstmann, L., de Oliveira, D., and Mattoso, M. (2024). Dlprov: A data-centric support for deep learning workflow analyses. In Proc. of the DEEM@SIGMOD 2024, page 77–85, New York, NY, USA. ACM.
Pina, D. B., Kunstmann, L. N. O., Chapman, A., de Oliveira, D., and Mattoso, M. (2025). Dlprov: a suite of provenance services for deep learning workflow analyses. PeerJ Comput. Sci., 11:e2985.
Siddique, M. A., Zaman, A., and Morimoto, Y. (2018). Selection of important sets by using k-skyband query for sets. International Journal of Advanced Computer Science and Applications, 9(4).
Talaei Khoei, T., Ould Slimane, H., and Kaabouch, N. (2023). Deep learning: systematic review, models, challenges, and research directions. Neural Computing and Applications, 35(31):23103–23124.
Venkataraman, S., Yang, Z., Franklin, M., Recht, B., and Stoica, I. (2016). Ernest: efficient performance prediction for large-scale advanced analytics. In Proceedings of the 13th Usenix Conference on Networked Systems Design and Implementation, NSDI’16, page 363–378, USA. USENIX Association.
Yao, S., Chen, H., Hanjie, A. W., Yang, R., and Narasimhan, K. (2023). Collie: Systematic construction of constrained text generation tasks.
Zhao, W. X. et al. (2023). A survey of large language models. arXiv:2303.18223, 1(2).
Publicado
08/09/2026
Como Citar
SANTOS, Gustavo; BEDO, Marcos; FROTA, Yuri; OLIVEIRA, Daniel.
Definição de Acordos de Nível de Serviço Personalizados para Treinamento de Modelos via Consultas k-Skyband. In: SIMPÓSIO BRASILEIRO DE BANCO DE DADOS (SBBD), 41. , 2026, São Carlos/SP.
Anais [...].
Porto Alegre: Sociedade Brasileira de Computação,
2026
.
p. 43-56.
ISSN 2763-8979.
DOI: https://doi.org/10.5753/sbbd.2026.249138.
