{"id":1873,"date":"2013-10-16T06:00:00","date_gmt":"2013-10-16T13:00:00","guid":{"rendered":"https:\/\/blogs.technet.microsoft.com\/dataplatforminsider\/2013\/10\/16\/sql-server-2014-pushing-the-boundaries-of-in-memory-performance\/"},"modified":"2024-01-22T22:49:08","modified_gmt":"2024-01-23T06:49:08","slug":"sql-server-2014-pushing-the-boundaries-of-in-memory-performance","status":"publish","type":"post","link":"https:\/\/www.microsoft.com\/en-us\/sql-server\/blog\/2013\/10\/16\/sql-server-2014-pushing-the-boundaries-of-in-memory-performance\/","title":{"rendered":"SQL Server 2014: Pushing the Boundaries of In-Memory Performance"},"content":{"rendered":"
This morning, during my keynote at the Professional Association of SQL Server (PASS) Summit<\/a> 2013, I discussed how customers are pushing the boundaries <\/em>of what\u2019s possible for businesses today using the advanced technologies in our data platform. It was my pleasure to announce the second Community Technology Preview (CTP2) of SQL Server 2014<\/a> which features breakthrough performance with In-Memory OLTP and simplified backup and disaster recovery in Windows Azure.<\/p>\n Pushing the boundaries<\/strong><\/p>\n We are pushing the boundaries of our data platform with breakthrough performance, cloud capabilities and the pace of delivery to our customers. Last year at PASS Summit, we announced our In-Memory OLTP project \u201cHekaton\u201d and since then released SQL Server 2012 Parallel Data Warehouse<\/a> and public previews of Windows Azure HDInsight and Power BI for Office 365<\/a>. Today we have SQL Server 2014 CTP2, our public and production-ready release shipping a mere 18 months after SQL Server 2012.<\/p>\n Our drive to push the boundaries comes from recognizing that the world around data is changing.<\/p>\n In-memory in the box for breakthrough performance<\/strong><\/p>\n A few weeks ago, one of our competitors announced plans to build an in-memory column store into their database product some day in the future. We shipped similar technology two years ago<\/em> in SQL Server 2012, and have continued to advance that technology in SQL Server 2012 Parallel Data Warehouse and now with SQL Server 2014. In addition to our in-memory columnar support in SQL Server 2014, we are also pushing the boundaries of performance with in-memory online transaction processing (OLTP). A year ago we announced project \u201cHekaton,\u201d and today we have customers realizing performance gains of up to 30x.\u00a0This work, combined with our early investments in Analysis Services and Excel, means Microsoft is delivering the most complete in-memory capabilities for all data workloads \u2013 analytics, data warehousing and OLTP.<\/p>\n We do this to allow our customers to make breakthroughs for their businesses. SQL Server is enabling them to rethink how they can accelerate and exceed the speed of their business.<\/p>\n A closer look into In-memory OLTP<\/strong><\/p>\n Previously, I wrote about<\/a> the journey of the in-memory OLTP project Hekaton<\/a>, where a group of SQL Server database engineers collaborated with Microsoft Research<\/a>. Changes in the ratios between CPU performance, IO latencies and bandwidth, cache and memory sizes as well as innovations in networking and storage were changing assumptions and design for the next generation of data processing products.\u00a0This gave us the opening to push the boundaries of what we could engineer without the constraints that existed when relational databases were first built many years ago.<\/p>\n Challenging those assumptions, we engineered for dramatically changing latencies and throughput for so-called \u201chot\u201d transactional tables in the database. Lock-free, row-versioning data structures and compiling T-SQL and queries into native code, combined with making the programming semantics consistent with SQL Server means our customers can apply the performance benefits of extreme transaction processing without application rewrites or the adoption of entirely new products.<\/p>\n\n
<\/a><\/p>\n\n