Kai Mast, A. Arpaci-Dusseau, Remzi H. Arpaci-Dusseau
{"title":"LambdaObjects","authors":"Kai Mast, A. Arpaci-Dusseau, Remzi H. Arpaci-Dusseau","doi":"10.1145/3538643.3539751","DOIUrl":null,"url":null,"abstract":"Existing cloud computing (or serverless) architectures provide convenient abstractions for application developers, but exhibit high latencies and do not support strong consistency guarantees. These limitations stem not only from the overheads due to virtualization, but also because storage and compute layers are disaggregated. We introduce LAMBDAOBJECTS; a new abstraction for server-less systems where data and compute are co-located. Data is encapsulated into objects, each associated with a set of methods that allow accessing and modifying the data. Functions then execute directly at the nodes the data is stored at. Our early results demonstrate that this architecture provides lower latencies, while sacrificing some elasticity offered by conventional serverless systems. In addition, LAMBDAOBJECTS provide strong consistency and enable efficient caching mechanisms and fault-tolerance with low overhead.","PeriodicalId":159120,"journal":{"name":"Proceedings of the 14th ACM Workshop on Hot Topics in Storage and File Systems","volume":"25 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2022-06-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"3","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Proceedings of the 14th ACM Workshop on Hot Topics in Storage and File Systems","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1145/3538643.3539751","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 3
Abstract
Existing cloud computing (or serverless) architectures provide convenient abstractions for application developers, but exhibit high latencies and do not support strong consistency guarantees. These limitations stem not only from the overheads due to virtualization, but also because storage and compute layers are disaggregated. We introduce LAMBDAOBJECTS; a new abstraction for server-less systems where data and compute are co-located. Data is encapsulated into objects, each associated with a set of methods that allow accessing and modifying the data. Functions then execute directly at the nodes the data is stored at. Our early results demonstrate that this architecture provides lower latencies, while sacrificing some elasticity offered by conventional serverless systems. In addition, LAMBDAOBJECTS provide strong consistency and enable efficient caching mechanisms and fault-tolerance with low overhead.