What we're about
Upcoming events (3)
We will explore Cosmos Db which is schema-less data storage. This includes exploring its storage structure, BSON document format and the MongoDb APIs. We will store, query and aggregate data using Python and .NET. We will compare traditional relational entity modeling to document modeling. We will also discuss sharding which is the partitioning method in Cosmos Db. I will demonstrate an IoT device collecting gyroscopic data and sending to Azure IoT Hub. An Azure function will store the data in Cosmos Db from service bus. The demonstration will utilize C# and PowerBI. Speaker Bio: Clarke Bowers is a software engineer with 35 years of experience. He has architected and developed embedded systems, desktop applications, enterprise data warehouses, web sites and web services; cloud bases and locally hosted solutions. He hold six patents and runs his own consulting business, details can be found here: http://www.cbsoftwareengineering.com/. The first database he created used Pascal on a PDP11 and he had to write the binary tree indexing code himself. Tables were limited to 65,536 records because that was the largest address possible on the 16-bit computer.
Microsoft made many wonderful changes to SQL Server with the release of SQL Server 2005, especially when it comes to querying LOBs (MAX, XML, and other large data-types). They also made a not-so-well advertised change that seriously affects the way that LOBs are handled and stored by default. Most people don’t even know this change occurred and, for those that do, most are unaware of the ramifications and rather extreme collateral damage this seemingly innocuous change has caused. 1. Non-Lob (which is most of our queries) Clustered Index Queries run twice as slow and require two orders of magnitude more memory to do so. 2. Rampant “bad” page splits, which can lead to serious blocking. 3. Permanent and perpetual fragmentation of Clustered Indexes. 4. Increased data storage requirements. 5. Increased and totally unnecessary Index Maintenance. 6. Increased log file activity, which also affects query performance, backup storage requirements, and increased backup and restore times. Fixing any one of those problems would be a big help. In this session, SQL Server MVP Veteran Jeff Moden shows us how and why the change causes all of these problems and then demonstrates how two simple changes to our tables fix it all. He also demonstrates how the same techniques can be used to make some non-LOB tables “Defragmented by Default”. Speaker Bio: Jeff Moden With more than 48,000 posts and 3 dozen mostly 5 star articles, Jeff Moden is a strong contributor on SQLServerCentral.com where he coined the term “RBAR” (Row By Agonizing Row) and helped make the "Tally Table" a household name. Jeff has more than 2 decades of experience with SQL Server and is mostly self-trained in what he calls the “Black Arts” of T-SQL. He’s known worldwide for his informative articles, high performance T-SQL coding methods, and methods of mentoring. His dedication to helping others earned him the MS SQL Server MVP award for nine years and the RedGate Exceptional DBA Award in 2011. His mantra is "Performance is in the code".
Details coming soon ...