Latest database related announcements from major vendors like Oracle and IBM
enable radical shifts in the way we design OLTP/OLAP databases and
applications. It is now possible to serve both OLTP and analytical needs from
the same database. Traditional database and application design processes
should be modernized to fully utilize new capabilities.
Contemporary corporate databases are roughly classified as either OLTP or
OLAP - analytical ( Data Marts, Data Warehouses ).
OLTP physical database design implies tables tied together via relationships
to describe business process. Historical information is not preserved. Year
end processes in GL, AP, AR applications, for example, typically close
current and roll over into next year, thus losing historical context.
OLAP databases, on the other hand, fully preserve historical information. It
is possible to analyze busine... (more)
Big Data and its most prominent technical ingredient, Machine Learning, are
all the rage these days, as IT industry is trying to convince companies
technology revolution is underway. ( "If you are not doing it, your
competitors sure are, and by the time you realize it, it will be too late" ).
Data fracking, i.e. Big Data, is 21 century new oil of that will power and
grease stalled industries and reignite growth, or so the story goes.
While advanced analytics (it comes under various names - predictive
analytics, data mining, and data science, more recently) is great and in use
(Please refer to the following article: Oracle 12c In-Memory, Columnar
Database & How It Relates to SAP Hana for update on IMDB/Columnar databases)
Contemporary large servers are routinely configured with 2TB of RAM. It is
thus possible to fit an entire average size OLTP database in memory directly
accessible by CPU. There is a long history of academic research on how to
best utilize relatively abundant computer memory. This research is becoming
increasingly relevant as databases serving business applications are heading
towards memory centric design and implementation.
If you si... (more)
All major relational database vendors are developing or already shipping
in-memory, columnar databases.
The next release of Oracle 12c - an in-memory, columnar database, will be
available next year. It will feature simultaneous transaction-level updates
to both row and column stores i.e. data will be stored in both formats at
the same time, in the same transaction. This is quite an improvement over SAP
Hana's awfully clumsy delta merge process ( data changes in SAP Hana are
first accumulated in delta store, then periodically merged into column store
- process which locks targe... (more)
Oracle 12c database and related suite of products are just released.
Columnar, compressed, high-speed, in-memory database to directly compete with
SAP Hana is what is promised in the next release of Oracle 12c database.
What we got so far is the familiar giant hairball that grew bigger.
In order to currently achieve Oracle's version of the holy grail of
cloud, in-memory, columnar, compressed, clustered services you must purchase,
integrate and manage a variety of Oracle products: 12c database, Exadata
software and hardware, TimesTen database, Oracle Clusterware, ASM etc.