This 3-part advanced course teaches the student
what the best (and worst) practices in NonStop Database (both ENSCRIBE and NonStop SQL/MP)
utilizations are. Many experienced people generalize database characteristics across platforms.
Tandem's databases have unique and powerful implementations that make fully
understanding them paramount in optimizing their performance.
How does DP2 really work? How do you make caching work for you? Is there one "best" file type?
When should we normalize our data, when shoud we de-normalize it? A relative file on
IBM works the same way as on Tandem...doesn't it?
If the wrong
decision can cripple performance, the right decision can make it sparkle!
This course is designed for senior Programmers, Analysts, Application Designers,
DBAs, System Managers and those who want to fully understand the unique and very powerful capabilities of
Tandem databases. It will cover Entity Design, Database Normalization, Caching, all three of
Tandem's file structures, plus unstructured files, and their respective strong (and weak) points.
Online: Self-Paced, includes Instructor Support
On-Site: 7 days
Those senior Programmers, Application Designers, Analysts, DBAs and System Managers and others
who need to learn about Tandem's file structures, DP2 process, Caching and other aspects of
Tandem databases.
- Logical Database Design
- Database Normalization
- DP2 Review
- Overview of General NSK File Strucutres
- Physical Database Design
- NonStop-specific aspects of Physical Database Design
- Review of DP2 structure and functions
- In-depth coverage of the three Tandem Strucutred file types
- In-depth coverage of how Tandem file types compare to others in the industry,
their relative strengths and weaknesses
- Minimum of 2 years experience on Tandem NonStop systems.
- Minimum 2 years hands-on experience with NonStop Databases (ENSCRIBE and/or NonStop SQL/MP).
|
- Logical Database Design
- Basic Components
- Logical Relational Design
- Database Normalization
- When to Normalize
- When to De-Normalize
- Normalization up to 5th Normal Form, when should you stop?
- Design Methodolgy
- Entity Relationships
- Methodologies
- ER Diagrams
- Relationships
- Unary, Binary, Tenary and N-ary Relationships
- Relationships as tables
- User Involvement
- Design best and worst practices
- Physical Database Design
- Best and worst practices
- To Index, or not to Index
- Platform Specifics
- Physical keys, and Index overhead
- Index-only scans
- MDAM, key design parameters, and performance
- Partitioning strategies and limitations; examples and methods
- Views, usages, and values
- General NSK File Structures
- General File Systems, and the NSK File System
- NSK Structured and Unstructured Files
- Why different file types
- Indexing and Partitioning
- File Extents
- Big Files (Format2 files)
- DP2 Review
- Disk Access Manager
- The DP2 Process
- Disk Cache
- Sequential Block Buffering
- File Extents
- Big Files (Format2 files)
- NSK File Infrastructure
- General NSK File Structures
- General File Systems, and the NSK File System
- NSK Structured and Unstructured Files
- Why different file types
- Indexing and Partitioning
- File Extents
- Big Files (Format2 files)
- OSS Files (Veritas)
- DP2 Review
- Disk Access Manager
- The DP2 Process
- Disk Cache
- Sequential Block Buffering (Real and Virtual)
- File Extents
- Big Files (Format2 files)
- NSK File-blocking strategy, including discussion of audited files and the
impact of auditing (TMF) on I/O performance.
- Detailed analysis and comparison of Relative, Entry-Sequenced and
Key-Sequenced files with similar file structures in the rest of the Industry
- Entry Sequenced Files - Compare and contrast to entry-sequenced formats
on other platforms. Discuss inserts, reads, updates, and alternate indexes.
- Relative files - Compare and contrast to other Relative file formats on other
platforms. We will also discuss space allocation and space reclamation (and
why NonStop excels at this), alternate keys, and use of three different types
of record insertion and the impact on key structure, including using logical
keys. Also discuss random vs. sequential reads - direct or with alternate indexes.
- Key Sequenced file. Compare and contrast this popular format with other Key-Sequenced
files on other platforms. We will discuss its major areas of strength on the NonStop
platform, and optimal utilization. Discussion also includes space allocation, reloading
strategies, and index structures as used in various platforms, including "fat index/
skinny data" designs, DB2 and NonStop. We will discuss indexed nonsequential strucutres
along with indexed sequential ones.
- Detailed design and structures of NSK Relative, Entry-Sequenced
and Key-Sequenced files,
- Entry-Sequenced Files - Discussion of NSK-Specific record alignment within blocks, free space, record
insertion, updates and deletes. We will discuss the Syskey for ES files, absence of block splits,
and optimal usage conditions.
- Relative Files - Discuss block allocation, record insertion and nature of the Syskey.
We will also discuss optimal usage conditions.
- Key Sequenced Files - Discuss specifics of block splits, record insertions, and deletions.
We will also discuss B-Tree index structure, impact of multiple index levels, when a Syskey
is or is not generated, caching strategies, and use of the Tandem Reload Analyzer (TRA).
|
