Creating Memory-Optimized Tables & Indexes

Memory-Optimized Tables

•       Defined as C structs, compiled into DLLs, and loaded into memory
•       Can be persisted as filestreams, or non-durable
•       Do not apply any locking semantics
•       Can be indexed using hash indexes
•       Can co-exist with disk-based tables
•       Can be queried using Transact-SQL
•       Cannot include some data types, including text, image, and nvarchar(max)
•       Do not support identity columns or foreign key constraints

Memory-Optimized Table Scenarios

•       Optimistic concurrency optimizes latch-bound workloads:
–      Multiple concurrent transactions modify large numbers of rows
–      A table contains “hot” pages
•       Applications should handle conflict errors:
–      Write conflicts
–      Repeatable read validation failures
–      Serializable validation failures
–      Commit dependency failures

Creating Memory-Optimized Tables

•       Add a filegroup for memory-optimized data

ALTER DATABASE MyDB
ADD FILEGROUP mem_data CONTAINS MEMORY_OPTIMIZED_DATA;
GO
ALTER DATABASE MyDB
ADD FILE (NAME = 'MemData' FILENAME = 'D:\Data\MyDB_MemData.ndf')
TO FILEGROUP mem_data;

•       Create a memory-optimized table

CREATE TABLE dbo.MemoryTable
(OrderId INTEGER NOT NULL PRIMARY KEY NONCLUSTERED HASH WITH (BUCKET_COUNT = 1000),
 OrderDate DATETIME NOT NULL,
 ProductCode INTEGER NULL,
 Quantity INTEGER NULL)
WITH (MEMORY_OPTIMIZED = ON, DURABILITY = SCHEMA_AND_DATA);

Memory-Optimized Indexes

•       Hash Indexes
–      Rows assigned to buckets based on hashed key
–      Multiple rows in the same bucket form a linked list
–      Effective for equality predicates
–      Query results are not sorted

CREATE TABLE tab1
(col1 INT NOT NULL INDEX h_idx NONCLUSTERED HASH
       WITH (BUCKET_COUNT = 100))
WITH (MEMORY_OPTIMIZED = ON,
      DURABILITY = SCHEMA_ONLY)

•       Range Indexes
–      Latch free, in-memory B-Tree structure
–      Effective for range scans, equality predicates, and inequality predicates

CREATE TABLE tab2
(col1 INT NOT NULL INDEX r_idx NONCLUSTERED)
WITH (MEMORY_OPTIMIZED = ON,
      DURABILITY = SCHEMA_ONLY)

Querying Memory-Optimized Tables

•       Query Interop
–      Interpreted Transact-SQL
–      Enables queries that combine memory-optimized and disk-based tables

•       Native Compilation
–      Stored procedure converted to C and compiled
–      Access to memory-optimized tables only







Comments

Post a Comment

Popular posts from this blog

SSRS Report Design: Best Practices

Introduction Microsoft® SQL Server™ Reporting Services is a comprehensive reporting tool. In using Reporting Services, you can design and deliver a wide range of reports, from simple tables to online free-form reports with graphical items and interactive links. Reporting Services can render reports for a variety of client-viewing formats, including HTML, PDF, Excel, and image formats. The wide variety of options and choices can be problematic for report authors who must make numerous decisions when choosing how to lay out a report. This paper provides a few guidelines to make these decisions easier and answers common questions.
Read more

SSAS OLAP Design - Best Practices

Introduction This paper outlines the recommended best practices for designing OLAP databases in Microsoft SQL Server 2005 Analysis Services to better meet the functional and performance needs of users. Designing a database for Online Analytical Processing (OLAP) databases is very different from designing a database for Online Transaction Processing (OLTP) and requires a different approach for achieving optimal results. The best practices recommendations are grouped into the following sections: Data Source Design Best Practices Dimension Design Best Practices Cube Design Best Practices Partition Design Best Practices Aggregation Design Best Practices Within each section, these best practices are ordered roughly by priority, where priority is a combination of how important the best practice is and how often you must worry about it. Major topics not covered include best practices for the design of calculations and for server management and operations.  These two topics are s...
Read more

Enable Usage-Based Optimization in SSAS

Usage-Based Optimization is a fantastic feature in SSAS to improve the most frequently used MDX queries of any cube database. This feature is used to design aggregations based on actual user queries using Business Intelligence Development Studio. But when ever you try to design  Usage-Based Optimization  you might have seen all the options  disabled.  This uses the technique of creating aggregations on FREQUENTLY USED queries in the given time interval.  We can enable the options in Usage-Based Optimization wizard. Connect to SSAS using SSMS (Sql Server Management Studio). Right click on the server and select properties as shown below. set CreateQueryLogTable to  true. Set the value of  QueryLogConnectionString  to any database where you wish to store your queries. Set the value of  QueryLogSampling  to the appropriate value for your environment (the default value of 10 means every 10th query will be logged). The lowe...
Read more