Microsoft Data Warehousing Offerings
When implementing a database solution, a “one shoe fits all” approach assumes that all database workloads are largely similar. Microsoft recognizes the different, sometimes conflicting, needs of targeted database workloads such as Online Transactional Processing (OLTP), Data Warehousing, and Business Intelligence.
Microsoft Data Warehousing offers a dedicated, optimized data warehousing implementation through a variety of offerings in Reference Architectures using SMP Architecture, and appliance based solutions scaling from entry-level data warehousing through Business Data Warehouse appliance to Massive Parallel Processing (MPP) offering maximal scale out solutions with SQL Server Parallel Data Warehouse.
Fast Track Data Warehouse Reference Architectures offer the customer a variety of choice in hardware implementations, optimized for data warehousing workload, providing scale, hardware optimization and flexibility.
Business Data Warehouse Appliance offers a balanced architecture, providing entry-level data warehousing with the shortest time to solution, in a single appliance based on the Fast Track Data Warehouse Reference architecture optimization and guidance.
SQL Server Parallel Data Warehouse provides the customer with a maximum scale, high performance massive parallel processing solution using an appliance that scales to multiple data and storage racks, in which each of the nodes that participate in the MPP architecture provide CPU capacity, storage and query optimization.
The Enterprise Data Warehouse appliance offers the highest data throughput, satisfying data warehousing needs from 10’s to 100’s of terabytes.
Which solution fits best?
When determining the correct solution for the customer, time to solution, query complexity, storage capacity, and concurrency are the key factors determining the right solution. MPP architecture offers a faster data load process over SMP architecture during high concurrent querying of existing data.
Lower entry data warehouses are a perfect candidate for the BDW appliance, based on the Fast Track Reference Architecture.
Separate database workloads are defined by the core functionality of applications, query patterns, and CPU and hardware utilizations.
These workloads can generally be divided into 3 separate, distinct and different workloads.
Transactional processing simplifies and accelerates data capture for accurate business decisions.
A transactional processing workload is defined by the volume of incoming transactions, the high-availability and data recovery needs associated with it, and concurrency depending on the amount of users simultaneously loading data.
Business insights are based on analysis of data that leads to optimized business processes and improved business performance. A business Insights solution is defined by data visualization through Business Intelligence tools. Microsoft is market leader in integrated business intelligence solutions, that are implemented through products as Sharepoint 2010, PerformancePoint Services, Office 2010, Power Pivot and SQL Server Analysis Services and SQL Server Reporting Services.
Data Warehousing enables a common data model to provide a single version of the truth, with data stored in an Enterprise Data Warehouse that offers choice, flexibility and storage capacity and query performance based upon the customer’s workload, defined by the complexity in query processing, concurrency and the data load process.
The Microsoft Data Warehouse offering provides a solid offering with different levels and stages of implementation providing the customer with specific optimizations defined by dedicated hardware and storage solutions as well as an optimized time to solution through reduced complexity.
The Data Warehousing offering is fitted towards the customer’s need with capability of growth (scale-up) and mass expansion (scale-out)
Within the Microsoft Data Warehousing offering, we offer different levels and stages of implementation, offering the customer choice, flexibility, and a solution that is fitted towards the customer’s need with capability of growth (scale-up) and mass expansion (scale-out).
Microsoft Data Warehousing Offerings
Microsoft Data Warehousing consists of box, reference architectures, and appliances dedicated to a specific workload, implementing SMP or MPP architecture.
SQL Server 2008 R2 Enterprise Edition
SQL Server Enterprise Edition offers a solid data warehousing solution, based on an out of box solution, allowing the customer the most flexible choice in hardware, storage or implementation. The drawback of a boxed product solution is defined by the manual installation and fine tuning towards a specific customer workload, resulting a longer time to solution based on the complexity of the implementation.
SQL Server Fast Track Data Warehouse – Reference Architecture
The Fast Track Data Warehouse Reference Architecture defines guidance and best practices on implementation, aligned with a balanced hardware architecture optimized with several hardware vendors and commodity hardware. Reference Architectures offer the best price performance for data warehousing that are ideal for data marts or small to mid-sized data warehousing.
The Reference Architecture based solutions offer variety in optimized hardware and choice of vendor, with scale up capabilities ranging in the 5 – 80 terabytes range.
SQL Server Parallel Data Warehouse – Enterprise Data Warehouse Appliance
The PDW (Parallel Data Warehouse) appliance offers high-end massive parallel processing at the highest performance and scale, in a fully integrated hardware and software solution with built-in high availability. SQL Server Parallel Data Warehouse offers unique ultra-shared nothing architecture in which tables can be organized in a distributed and replicated table model, offering maximal query throughput, and high-intense load processes without interfering with a high level of concurrent query and CPU workloads. The Architecture consists of a Control Rack that takes control over client connectivity, appliance management and load processes, with individual query and data retrieval sent to a data rack that consists of each of the individual compute and storage nodes that make up the Massive Parallel Processing architecture. This high-end enterprise data warehousing solution offers maximum scale and expansion allowing additions of data racks to maximize scale out, functioning as a hub in distributed data warehouse architecture (hub/spoke).
Teched North America – All UP Data Warehouse – From SMP to Parallel Data Warehouse
DevDays 2011 Netherlands – Deep Dive into Microsoft Data Warehouse