{"title":"通过对主存数据库的散列预测处理多连接查询的成本","authors":"Feilong Liu, Spyros Blanas","doi":"10.1145/2806777.2806944","DOIUrl":null,"url":null,"abstract":"Database management systems (DBMSs) carefully optimize complex multi-join queries to avoid expensive disk I/O. As servers today feature tens or hundreds of gigabytes of RAM, a significant fraction of many analytic databases becomes memory-resident. Even after careful tuning for an in-memory environment, a linear disk I/O model such as the one implemented in PostgreSQL may make query response time predictions that are up to 2× slower than the optimal multi-join query plan over memory-resident data. This paper introduces a memory I/O cost model to identify good evaluation strategies for complex query plans with multiple hash-based equi-joins over memory-resident data. The proposed cost model is carefully validated for accuracy using three different systems, including an Amazon EC2 instance, to control for hardware-specific differences. Prior work in parallel query evaluation has advocated right-deep and bushy trees for multi-join queries due to their greater parallelization and pipelining potential. A surprising finding is that the conventional wisdom from shared-nothing disk-based systems does not directly apply to the modern shared-everything memory hierarchy. As corroborated by our model, the performance gap between the optimal left-deep and right-deep query plan can grow to about 10× as the number of joins in the query increases.","PeriodicalId":275158,"journal":{"name":"Proceedings of the Sixth ACM Symposium on Cloud Computing","volume":"21 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2015-07-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"23","resultStr":"{\"title\":\"Forecasting the cost of processing multi-join queries via hashing for main-memory databases\",\"authors\":\"Feilong Liu, Spyros Blanas\",\"doi\":\"10.1145/2806777.2806944\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"Database management systems (DBMSs) carefully optimize complex multi-join queries to avoid expensive disk I/O. As servers today feature tens or hundreds of gigabytes of RAM, a significant fraction of many analytic databases becomes memory-resident. Even after careful tuning for an in-memory environment, a linear disk I/O model such as the one implemented in PostgreSQL may make query response time predictions that are up to 2× slower than the optimal multi-join query plan over memory-resident data. This paper introduces a memory I/O cost model to identify good evaluation strategies for complex query plans with multiple hash-based equi-joins over memory-resident data. The proposed cost model is carefully validated for accuracy using three different systems, including an Amazon EC2 instance, to control for hardware-specific differences. Prior work in parallel query evaluation has advocated right-deep and bushy trees for multi-join queries due to their greater parallelization and pipelining potential. A surprising finding is that the conventional wisdom from shared-nothing disk-based systems does not directly apply to the modern shared-everything memory hierarchy. As corroborated by our model, the performance gap between the optimal left-deep and right-deep query plan can grow to about 10× as the number of joins in the query increases.\",\"PeriodicalId\":275158,\"journal\":{\"name\":\"Proceedings of the Sixth ACM Symposium on Cloud Computing\",\"volume\":\"21 1\",\"pages\":\"0\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2015-07-11\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"23\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Proceedings of the Sixth ACM Symposium on Cloud Computing\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1145/2806777.2806944\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Proceedings of the Sixth ACM Symposium on Cloud Computing","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1145/2806777.2806944","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
Forecasting the cost of processing multi-join queries via hashing for main-memory databases
Database management systems (DBMSs) carefully optimize complex multi-join queries to avoid expensive disk I/O. As servers today feature tens or hundreds of gigabytes of RAM, a significant fraction of many analytic databases becomes memory-resident. Even after careful tuning for an in-memory environment, a linear disk I/O model such as the one implemented in PostgreSQL may make query response time predictions that are up to 2× slower than the optimal multi-join query plan over memory-resident data. This paper introduces a memory I/O cost model to identify good evaluation strategies for complex query plans with multiple hash-based equi-joins over memory-resident data. The proposed cost model is carefully validated for accuracy using three different systems, including an Amazon EC2 instance, to control for hardware-specific differences. Prior work in parallel query evaluation has advocated right-deep and bushy trees for multi-join queries due to their greater parallelization and pipelining potential. A surprising finding is that the conventional wisdom from shared-nothing disk-based systems does not directly apply to the modern shared-everything memory hierarchy. As corroborated by our model, the performance gap between the optimal left-deep and right-deep query plan can grow to about 10× as the number of joins in the query increases.