"We were initially very skeptical that AutoTMF would work in our extreme
processing environment or with our custom software. After thorough
stress testing, the decision to move forward using AutoTMF (and RDF) was
a simple one which we have never had any reason to regret."
VP Development, Solutions Provider
"While our initial intent for using AutoTMF was to enable reliable
disaster recovery, we are jazzed about the performance improvements we
have seen as a result of implementing TMF. I can't believe we waited all
these years to implement TMF and thanks to AutoTMF, we didn't have to
change our programs to do it."
General Manager, Service Bureau
NonStop AutoTMF provides automatic data protection and
improved performance for your applications by implementing NonStop TMF
on behalf of your programs.
No change to your existing programs is required.
Sold and supported worldwide by HP.
AutoTMF automatically implements TMF transactions on behalf of your programs where necessary for
both Enscribe files and NonStop SQL tables managed by Escort SQL.
What TMF/RDF Does For You
If your database is audited, NonStop TMF guarantees that every file update has an audit trail record. In conjunction with TMF online dumps, the audit trail can be used to recover lost files up to the state of the last committed update.
For an excellent description of NonStop TMF, the article 'a
conversation with
NonStop TMF architect Charles Johnson' is an excellent source.
Once the databases are TMF protected,
RDF uses the TMF audit trails to reliably and efficiently replicate the database to a second copy on a remote
NonStop system. RDF guarantees that all audited updates are quickly transmitted to the remote system and applied to the remote copy of
the database.
For serious NonStop-to-NonStop replication-and for proven value in disaster recovery situations-there is no more efficient, reliable, or complete solution than TMF/RDF.
What NonStop AutoTMF Does For You
Works with all applications
NonStop
AutoTMF allows all applications to use TMF for file/volume recovery and RDF for disaster recovery and data replication, even if they are not written to use TMF.
Once enabled on your system, AutoTMF dynamically determines effective and efficient transaction boundaries and automatically causes your application to make audited updates to the database.
There is no need to make any source code changes to your programs, or even to recompile them. AutoTMF works with your existing application object files.
AutoTMF makes the migration to TMF protected data and programs, smooth, fast, efficient and low risk. You can convert one file to TMF auditing or convert them all!
Improves Application Performance
The NonStop Kernel DP2 disk process has optimized performance for audited files. Its caching algorithms assume that database updates can be reliably buffered without requiring an individual disk write. Updates are collected and written to the audit trail to provide complete reliability in the event of system failure.
Another big performance booster can be the elimination of block-splitting
overhead as a result of TMF auditing of files. In some cases, performance gains
from auditing files subject to block-splitting has been truly impressive. For an
example, see the 2003 ITUG presentation from one user:
Performance Characteristics of Audited and
Non-Audited Files
In most cases, audited files result in far lower I/O activity. They also tend to eliminate update hotspots for log files and random updates.
Application benchmarks of AutoTMF show no reduction in throughput
and response time. In many cases, it has completely eliminated I/O hotspots
and improved performance.
Using AutoTMF, applications gain all the benefits of TMF, without being required to include the transactional coding, in your programs.
Enables incremental migration to TMF
NonStop
AutoTMF makes use of advanced technology to determine optimum transaction boundaries for performance. Although AutoTMF will automatically enable your database for use with TMF and RDF with no change to your programs, no automatic method can recognize the boundaries of your business transactions. To even further improve disaster recovery down to the business transaction level and further increase overall performance, you may place transaction calls in selected application programs.
A traditional "Big Bang" conversion to TMF would require all programs to be converted before the database can be audited. With AutoTMF, however, you can migrate incrementally-one program at a time.
By including TMF calls directly in your programs, you can more accurately control the transaction level atomicity and ensure that complete business transactions are either completely done or undone as a unit. AutoTMF provides transactions only for those programs that do not contain transaction calls. Should you include TMF calls in one or more programs, AutoTMF will cease providing transactions in those cases.
AutoTMF performs a onetime step for each program that adds the AutoTMF runtime library and makes a number of reference changes. The preparation step does not modify the program in any detectable way (i.e. source code, object code, timestamp, etc. all remain unchanged).
Disaster Recovery, Improved Data Consistency, and Improved Performance
NonStop
AutoTMF allows your company to instantly benefit from the high
performance and reliability of hp's NonStop TMF and RDF technologies, without having to modify your programs. Implementation is fast, simple, and low risk. Your data is the heart of your business -protect it with AutoTMF.
