![\"Landscape](\"/static/b0453fbb95c32586f445dca4f9795c72/3e992/storage-engine-tree.png\" "\\"Landscape")
\n This is the second article of a series to summarize the key concepts of Chapter 3. Storage and Retrieval in the Designing Data Intensive Application book. The series consists of 3 articles, including log-structure storage engine (SSTables and LSM Tree), page-oriented storage engine (B Tree) and column based storage engine.
\n\n In short, database storages are divided into two categories, row-based and column based. We will talk more about the difference of them in next article. In this article, we will talk more about the page-oriented storage under row-based storage.
\nAlmost log-structured storage engine have pro-found influence on database evolution, the most widely adopted engine is actually page-oriented storage engine.
\nDifferent from log-structure storage engine, page oriented storage engine update the value in place (in its original page), instead of appending data to latest segment file.
\nB Tree, which were adopted by MySQL, PostgreSQL and MongoDB, are the most adopted data structure.
\n![\"Landscape](\"/static/aceba302c4d7fedfdc314c15790959b5/11d70/b-tree.png\" "\\"Landscape")
\n For read request:
\nFor write request:
\nBranching factor: the number of nodes in one page, usually it’s several hundreds.
\n| Category | \nPros & Cons | \nWell know data structure | \nReal world DB | \n
|---|---|---|---|
| Log-structure | \nBetter write performance | \nSSTables & LSM Tree | \nCassandra LevelDB HBase Lucence | \n
| Page-oriented | \nBetter read performance more complex concurrency control & crash recovery | \nB Tree | \nMySQL PostgreSQL MongoDB | \n
Write Amplification: one write query can produce how many disk writes ?
\nLog structure engines have higher write latency at high percentiles:
\nIn this article, we walk through the design of page-oriented storage engine, mainly B Tree. We can know that data was updated directly in page file, so in general we can have worse write performance. But Since it’s a balanced tree structure, it can have much better read performance.
\nSo in short, if the scenario requires much more read than write, we can consider use page-oriented storage engine. On the other hands, if we to handle a write intensive application, we can consider using log-structured storage engine.
\nIn next article, we will talk more about column-based storage engine, which has many different design philosophy comparing to row-based storage engine. Let’s go.
","fields":{"slug":"/posts/btree","tagSlugs":["/tag/database-storage/","/tag/b-tree/","/tag/page-oriented-storage-engine/","/tag/designing-data-intensive-application/"]},"frontmatter":{"date":"2021-12-18T23:12:04.772Z","description":"B Tree is one of the most famous data structure of page-oriented storage engine (e.g. MySQL). page-oriented storage engine has relatively good read performance comparing to log-structured storage engine. Let's take a look at B Tree!","tags":["Database Storage","B Tree","Page oriented storage engine","Designing Data Intensive Application"],"title":"Storage and Retrieval (2) - B Tree","socialImage":{"publicURL":"/static/e1c2867b251a4f300016b135407f731f/social.jpg"}}}},"pageContext":{"slug":"/posts/btree"}},"staticQueryHashes":["251939775","401334301","825871152"]}