Madison --- Microsoft SQL Server 2008 R2 Parallel Data Warehouse

Another product from Microsoft that is going to revolutionize the IT industry is the Microsoft SQL Server 2008 R2 Parallel Data Warehouse. It was previously code-named  "Madison". In this post I am going to discuss about it and its benefits.

Introduction

Microsoft SQL Server 2008 R2 Parallel Data Warehouse (previously code named project "Madison") is a highly scalable appliance that delivers performance at low cost through a massively parallel processing (MPP).

The goal of Madison is to use MPP (massively parallel processing) to deliver high performance and scalability on SQL Server 2008, Windows Server 2008 and industry-standard hardware. The appliance partitions large tables across multiple physical nodes, with each node having dedicated CPU, memory and storage and running its own instance of SQL Server.


The query processing takes place within one physical instance of a database. A Control Node routes queries from applications to all Compute Nodes and then collects and returns the result. Since the data is evenly distributed across multiple nodes and processing occurs in parallel, queries can be performed faster than on single SMP (symmetric multiprocessing) database servers, according to Microsoft.

Microsoft has dubbed the product's parallel design "Ultra Shared Nothing."

Microsoft has made a number of announcements around its database and data warehousing business of late. According to Ron Van Zanten, directing officer of business intelligence, Premier Bankcard LLC, “Upgrading to SQL Server 2008 has improved the performance of our data warehouse significantly.  With more than 20 terabytes of data and growing, we are ready to take our data warehouse to the next level.  In our current evaluation of Madison we’re very impressed with the efficiency and raw power of the appliance and how it can really process many terabytes of data seamlessly.” Appliances ranging from 8 nodes to 20 nodes are now ready to host customers for early evaluation.

Benefits

Madison is a highly scalable data warehouse appliance that offers customers high performance at low cost through a massively parallel processing (MPP) architecture for SQL Server. Unlike its competitors, Madison offers hardware flexibility with configurations from the major hardware vendors and low cost through industry standard hardware. Project Madison extends Microsoft investments in data warehousing and allows SQL Server data warehouses to scale out, from fewer than 50 gigabytes to over a petabyte of data. It will utilize a Massively Parallel Processing (MPP) scale-out architecture to support data warehousing ranging from 10TB to 1+ PB. The Parallel Computing Edition is priced at $57,489 per processor.

Key customer benefits  “Madison” include:

• Data warehouse scalability into the petabyte range
• Low cost of ownership through industry standard hardware
• Appliance model that simplifies data warehouse deployment and maintenance
• Integrates with existing SQL Server 2008 data warehouses via a unique hub-and-spoke architecture

Key Features

• Data warehouse scalability from tens to hundreds of terabytes
• Low cost of ownership through industry standard hardware
• Appliance model simplifies deployment and maintenance
• Integrates with existing SQL Server 2008 data warehouses via hub-and-spoke architecture
• Greater ROI from BI investments through integration with SQL Server 2008
• Reduced risk through use of redundant, industry standard hardware
• Balanced reference architectures deliver predictable performance
• Better agility and business alignment through hub and spoke architecture

Scale out with Parallel Data Warehouse

A SQL Server 2008 R2 Parallel Data Warehouse MPP appliance acts as an enterprise “hub” that publishes data as needed to various business units or departments (spokes). For dedicated high performance requirements, individual business units can redeploy SQL Server 2008 data marts, or deploy their own Parallel Data Warehouse appliances, as spokes.

Scale Your Data Warehouse from tens to hundreds of Terabytes

SQL Server 2008 R2 Parallel Data Warehouse is a highly scalable appliance for enterprise data warehousing. It is the next step in the evolution of the data warehouse appliance created by DATAllegro. SQL Server 2008 R2 Parallel Data Warehouse uses massively parallel processing (MPP) to deliver the high performance and scalability on SQL Server 2008, Windows Server® 2008 and industry-standard hardware. The MPP architecture helps enable better scalability, better and more predictable performance, reduced risk and a lower cost per terabyte than other DW solutions.

Improve Performance at a Lower Price per Terabyte

In symmetric multi-processing (SMP) architecture, query processing occurs entirely within one physical instance of a database. CPU, memory and storage impose physical limits upon speed and scale. A SQL Server 2008 R2 Parallel Data Warehouse MPP data warehouse appliance partitions large tables across multiple physical nodes, each node having dedicated CPU, memory and storage, and each running its own instance of SQL Server, in a parallel shared nothing architecture. SQL Server 2008 R2 Parallel Data Warehouse delivers improved performance through its unique design.

All components are balanced against each other to reduce performance bottlenecks, and all server and storage components are mirrored for enterprise-class redundancy. A Control Node routes queries from applications to all Compute Nodes, then collects and returns the result. Because data is evenly distributed across multiple nodes and processing occurs in parallel, queries can be many times faster than on single SMP database servers.

This architecture can cost less for several reasons:

• Rather than relying on expensive proprietary processors or storage, SQL Server 2008 R2 Parallel Data Warehouse appliances use industry-standard hardware
• As data volumes grow, scalability simply requires the addition of capacity to the appliance
• There is no need for a “forklift upgrade” where the entire appliance must be upgraded

Improve Performance at a Lower Price per Terabyte

In a traditional, symmetric multi-processing (SMP) architecture, query processing occurs entirely within one physical instance of a database. CPU, memory and storage impose physical limits upon speed and scale.

A Madison MPP data warehouse appliance partitions large tables across multiple physical nodes, each node having dedicated CPU, memory and storage, and each running its own instance of SQL Server, in a patented parallel design known as Ultra Shared Nothing™. All components are balanced against each other to reduce performance bottlenecks, and all server and storage components are mirrored for enterprise-class redundancy.

A Control Node routes queries from applications to all Compute Nodes, then collects and returns the result. Because data is evenly distributed across multiple nodes and processing occurs in parallel, queries can run many times faster than on single SMP database servers.

This architecture can cost less because, rather than relying on expensive proprietary processors or storage, Madison appliances use industry-standard hardware. As data volumes grow, scalability simply requires the addition of capacity to the appliance. There is no need for a “forklift upgrade” where the entire appliance must be upgraded.

Improve ROI of Existing BI Investments

Extract more value from existing BI investments through tight integration of Madison and Microsoft BI tools. The Madison MPP data warehouse integrates with SQL Server technologies including Integration Services, Reporting Services and Analysis Services. Existing SMP data arts can be redeployed and “plugged in” to the MPP cluster as “spokes”,receiving data from the MPP “hub.”

Appliance Model

Simplify data warehouse deployment and maintenance. Madison appliances include carefully balanced, pre-assembled hardware and software from leading hardware vendors. This “appliance model” significantly accelerates your time to value and reduces deployment costs.

Deliver Predictable Performance

Through carefully balanced configuration, tight specifications and testing, the Madison MPP data warehouse appliance is designed to deliver predictable performance. Microsoft designs the reference configurations to suit different usage scenarios, including reporting and ad hoc queries.

Automated workload management and system resource balancing enable you to load data concurrently with queries, without sacrificing performance. As data grows and business needs change, Microsoft offers options to support more users performing different types of queries.

Better Agility and Business Alignment through Hub-and-Spoke

Using a true hub-and-spoke architecture, all enterprise data can be maintained on a Madison hub while departments or business units keep their existing data marts to suit their needs. High-speed data transfer relieves traditional barriers to hub and spoke. Power users can even deploy a dedicated MPP appliance as a “spoke” so they can autonomously manage resources, while IT can enforce enterprise standards across all data.

Additional Benefits

Automated storage management replaces complex space allocation. Available diagnostics detect and broadcast hardware issues. A management console provides a simple GUI interface and dashboard for tracking and resolving issues. Ultra-fast loading and high-speed backups reduce administration for even the largest data warehouses.


You can use Madison MPP data warehouse applicance as:

• As an economical, high-performance enterprise data warehouse or data mart
• As a platform for high scale Data Warehouses from terabytes to over 1 petabyte
• As the hub of a hub-and-spoke EDW
• To offload applications from an EDW for high-speed aggregation and queries
• As long-term data storage in multi-tier data warehouses
• As a subject-oriented data mart or sandbox for new analytic applications

Further Readings

For more information about Project “Madison”, visit http://www.microsoft.com/madison

A good video explaining Madison is at MSDN Channel9.

Do keep me updated with your views and updates.