Select Content Type
What kind of article do you want to create?
Close Window
Article
Schedule
ArticleID
54591 - Click to preview
Canonical URL
https://www.dbta.com/Columns/SQL-Server-Drill-Down/Reaching-for-Highly-Scalable-Systems-with-SQL-Server-2008-54591.aspx
Title
StandOut Url
Author(s)
Kevin Kline
Images
Related Articles
Summary
SQL Server has supported VLDBs (very large databases) for some time now. Back in the SQL Server 2000 days, I recall hearing multi-terabyte databases were unusual but doable. Now, they are commonplace, while databases in the hundreds of terabytes inhabit the part of the map that says "there be dragons." While VLDBs are quite common on SQL Server today, highly scalable systems that can be flexibly extended in the same fashion as Oracle/RAC are less so. So, how do you design a highly available architecture for SQL Server if it's not like Oracle/RAC.
Page 1
Page 2
Page 3
Page 4
Page 5
Page 6
Page 7
Page 8
Page 9
Page 10
Page 11
Page 12
Page 13
Page 14
Page 15
Video
<p>SQL Server has supported VLDBs (very large databases) for some time now. Back in the SQL Server 2000 days, I recall hearing multi-terabyte databases were unusual but doable. Now, they are commonplace, while databases in the hundreds of terabytes inhabit the part of the map that says "there be dragons." While VLDBs are quite common on SQL Server today, highly scalable systems that can be flexibly extended in the same fashion as Oracle/RAC are less so. So, how do you design a highly available architecture for SQL Server if it's not like Oracle/RAC.</p><p>Available third-party tools include xkoto's GRIDSCALE technology. Older architectures also can be implemented for a scaled-out application such as distributed partitioned views, or a two-phase commit using DTC (data transaction coordinator). Since those older technologies involve various costly tradeoffs in performance, maintainability, or implementation, and third-party products can be costly, what other options exist for a SQL Server application?</p><p>You can exploit some new features quickly and easily if you have a specific non-OLTP scenario requiring flexible scalability. For example, you can use the scalable shared database architecture, which requires a SAN and is described at <a href="http://support.microsoft.com/kb/910378" target="_blank">support.microsoft.com/kb/910378</a> , to build highly flexible and extensible hardware infrastructure for reporting applications and Business Intelligence (BI) data marts. The database must be in read-only mode most of the time, so its strongest applicability is to reporting and BI data marts.</p><p>You also can read white papers by the Microsoft SQL Server Customer Advisory Team (SQLCAT) that provide excellent techniques to scale out Analysis Services, with or without a SAN, at <a href="http://sqlcat.com/whitepapers/default.aspx" target="_blank">sqlcat.com/whitepapers/default.aspx</a> .</p><p>Through acquisition, Microsoft also is actively advancing its high-end architecture for database warehousing and BI. Microsoft recently acquired DATAllegro, a maker of data warehousing appliances, and, using DATAllegro's technology, Microsoft data warehouses can effectively scale to hundreds of terabytes, even reaching petabyte sizes.</p><p>Some analysts say the win for Microsoft is huge because the reach of its enterprise scalability now enters the rarefied air of the likes of Teradata, while others are not sure this is a sustainable, competitive differentiator for the long run. Another compelling aspect of the technology is its open architecture. You can select from a menu of hardware platforms, as well as the grid architecture, where you can distribute processing power and memory wherever it is needed. See the FAQ at <a href="http://www.datallegro.com/microsoft/microsoft_faq.asp" target="_blank">www.datallegro.com/microsoft/microsoft_faq.asp</a> .</p><p>As with any such acquisition, I suspect it will be 18-24 months before the software reaches a version 1.0 launch; so, working back from that date, early adopters should be working on the-product-formerly-known-as DATAllegro. The way it is released and packaged-whether as software, software plus hardware, or an entire turn-key appliance-remains to be seen.</p><p>It's clear, however, that DATAllegro software running on Windows Server 2008 and SQL Server 2008 will be capable of assaulting territory only Teradata could hold in the past. With Microsoft's cut-rate pricing policies, I'd expect this offering to cost a fraction of what Teradata charges. Also, since the product runs on inexpensive commodity hardware, even modestly budgeted IT departments will be able to afford high-end data marts with confidence, while maintaining the IT hardware purchasing standards of their favorite commodity vendors.</p><p>If you've been around for awhile, you'll recognize the pattern. Microsoft has repeatedly tackled the domain of a specific high-priced specialty market and made it a subset of the SQL Server product set. It happened in SQL Server 7.0 when Data Transformation Services first challenged high-cost ETL products. It happened in SQL Server 2000 when Reporting Services challenged expensive vendors like Crystal Reports. The SQL Server 2005 and Performance Point Server package is a strong challenger to Cognos. So, it looks like other vendors in the wider market should continue to glance over their shoulders.</p>
Newsletter Name
Issue Name
Article Type
Article SubType
DBTA E-Edition
January 2009
Columns
SQL Server Drill Down
Please Choose
5 Minute Briefing : Blockchain
5 Minute Briefing: Cloud
5MB: Data Center
5MB: Information Management
5MB: MultiValue
5MB: Oracle
5MB: SAP
Big Data Quarterly Issue
Cloud Strategies
DBTA E-Edition
ExaBriefing
Headlines from AIOUG
IBM LinuxLine
Infrastructure Wisdom
IOUG Storage Systems
Linux Executive Report from IBM
Magazine Issue
Oracle Enterprise Manager
Unisphere Five Minute Briefing
Columns
Editorial
A Wider View
Applications Insight
Big Data Notes
Database Elaborations
DBA Corner
Defining Data
Emerging Technologies
From 30,000 Feet
MongoDB Matters
My View
MySQL Musings
New Directions
Next-Gen Data Management
Notes on NoSQL
Oracle Data Strategies
Oracle Observations
Quest IOUG Database & Technology Insights
SQL Server Drill Down
The Enterprise Environment
The Open DBA
The Philosophy of PL/SQL
Trends and Observations
Categories
Topics
Artificial Intelligence
Big Data
Blockchain
Business Intelligence and Analytics
Cloud Computing
Data Center Management
Data Integration
Data Modeling
Data Quality
Data Warehousing
Database Management
Database Security
Hadoop
Internet of Things
Master Data Management
MultiValue Database Technology
NoSQL Central
Virtualization
×
Authors
×
Search Articles
×
Image Helper
Upload an Image
Name
File
Browse…
Image Resize (width):
90
120
135
250
No resize
Preview (Click image to select)
Select the Image
ID
Image Name