Getting Started with Adobe After Effects - Part 6: Motion Blur


Upload Image Close it
Select File

Explore the features being added with each release of SQL Azure

Moderators
Top Categories · View All
SU8(Dec 2011) 7
SU7(July 2011) 6
SU6(May 2011) 4
SU5 (Oct 2010) 4
SU2(APR 2010) 4
SU1(Feb 2010) 4
SU4(Aug 2010) 3
SU3(June 2010) 3

Browse by Tags · View All
Spatial Types 3
Availability 2
SQL Azure Management Portal 2
Database Size 2
Metro Styled UI 1
User Collations 1
Database cost 1
Federations 1
Performance 1
Scalabilty 1
Co-Administrator support 1
Firewall rules 1
DAC Upgrade 1
JDBC Driver 1
DB Management REST API 1
Multiple servers for subscription 1
sp_tableoption 1
Error Messages 1
Reporting 1
Sync Support 1

Browse by Release · View All
All 35

SQL Azure

SQL Azure federations gives scale-out performance

Mar 18 2012 12:00AM by Ramireddy   

One traditional approach to improve performance is scale-up by increasing hard-ware etc. Scale-up approach has certain limit. New trend towards improving performance is scale-out. SQL Azure supports scale-out by using federations. This will solve the problem of databases will have 150GB maxsize also.

In this approach data will be divided into multiple subsets and stored across multiple federations. Rows will be horizontally partitioned across federations. This approach is also called as "sharding". While creating federations, initially federation scheme has to be defined, this federation scheme defines a federation distribution key along with its lower and upper range values for each federation in which data has to be stored. When the row is inserted for federated tables, according to the key value, it will be inserted into corresponding federation.

Read More Have you written a blog post or article related to this feature and want to get them added here? Read More...
Category : SU8(Dec 2011)
Release : All
Tags : Federations

Ramireddy
2 · 41% · 12972
1


Submit

1  Comments  

  • Are you sure SQL Azure Federations are about performance? I believe they are more about scale.

    Horizontal distributed partitioning in an on-site implementation can be said to have a performance benefit due to the assumption that the additional hardware will be dedicated to the task at hand. It must be in order to overcome the network hit required to resolve queries where data lives on more than one partition.

    But I don't think the same assumptions apply to SQL Azure. You simply have no assurance that each shard will run on dedicated hardware. Quite the opposite, you should assume it's all running on shared servers.

    I've asked a number of people if there would be a benefit to doing this:

    Assumptions: 100 GB of data Even key distribution

    Option 1: One Azure instance holding 100 GB of data (50GB unused space available for growth) Option 2: Five Azure instances holding 20 GB of data each, Federated into a single 100 GB database

    I've been told by multiple people that Option 1 will perform better.

    Have you tested SQL Azure Federations for performance? I have not yet had the opportunity.

    commented on May 7 2012 4:45AM
    Marc Jellinek
    97 · 2% · 546

Your Comment
Sign Up or Login to post a comment.
"SQL Azure federations gives scale-out performance" rated 5 out of 5 by 1 readers
SQL Azure federations gives scale-out performance , 5.0 out of 5 based on 1 ratings
    Copyright © Rivera Informatic Private Ltd Contact us      Privacy Policy      Terms of use      Report Abuse      Advertising      [ZULU1097]