dbi services Blog
Welcome to the dbi services Blog! This blog focuses on IT infrastructure - featuring news, troubleshooting, and tips & tricks. It covers database, middleware, and OS technologies such as Oracle, Microsoft SQL Server, Documentum, MySQL, PostgreSQL, Sybase, Unix/Linux, etc. The dbi services blog represents the view of our consultants, not necessarily that of dbi services. Feel free to comment on the postings!
Oracle Open World 2013: First day - In-Memory, Perl for DBAs, and RDBMS Forensics
Oracle Open World 2013 just started this Sunday morning, 22 October. I'm very happy to be part of such an event this year as an attendee and as a speaker. It's a very impressive event with lots of people (60.000) until the 26th. In the following days, I'll keep you informed of the most interesting things I'll see.
Let's talk about the first day!
Even for the first day and for a Sunday, it was a very busy day: two interesting sessions, a presentation and the opening Keynote with some announcements from Larry Ellison.
I started with a session about Perl by Arjen Visser, CTO of Dbvisit:
- Perl: A DBA’s and Developer’s Best (Forgotten) Friend [UGF2182]
dbi services is official partner of Dbvisit in Switzerland and some parts of the software is written in Perl. It was a pretty interesting session because as an old developer I'm still interested in how we can create great software to help us in our work. The session gave us an overview aft a language that can still be very helpful because modules thanks to CPAN can enhance it easily.
Then I chose a session because of its title:
- RDBMS Forensics: Troubleshooting with Active Session History [UGF9740] by Tim Gorman
The term forensics made me think that we will do an expert job like in C.S.I (the TV show).
In that session, we've seen that the same error ORA-3135/ORA-3136 can have different root causes and totally different ways to solve it.
It confirms that Active Session History (ASH), which is mostly used as a tuning tool, can be a very powerful diagnostic tool.
I've hosted a session called "Solving Critical Customer Issues with the Oracle Database 12c Optimizer [UGF5498]" in the beginning of the afternoon in Moscone West.
It's interesting to share what I've found during my tests on Oracle 12c. The session was about 3 customer cases I reproduced on virtual machines and how the new Oracle version deals about theses issues. All scripts will be available on blogs, the first one has already been published.
Later in the afternoon, the opening Keynote took place, there were several parts but I will focus on the parts that impressed me much.
Noriyuki Toyoki, Corporate Senior Vice President, Fujitsu, introduced the concept of Big Data and the need of performance to handle Big Data and perform analysis in real-time. Handling Big data give an advantage on competitors and is a way of evolution for companies.
With a partnership with Oracle, the new SPARC processors from Fujitsu M10 will be more efficient because of database software embedded directly on chip, for example hash and cryptographic functions and so one are now directly in the processor and not done by software. It improves the efficiency.
Then Larry Elisson, CEO of Oracle, focused on the new In-memory database option. The new In-memory concept will store in memory not only the tables in rows as the standard way but also in columns. Processing data in columns is more efficient when reading data whereas storage in rows is more efficient to ensure the transactional activities (insert data for example).
According to Oracle, the new in-memory solution is at least 100 times better that traditional ways. A live demo showed more than 1300 times improvement on statement reading more than 200 billions of rows.
Larry also explained that the new option is 2 times faster on transactional features. How? Just be removing 'analytic' indexes that will become useless because of the column storage.
That new option should improve the performance of every application without any code change by just enabling an init parameter.
I hope I'll be able to test such feature because on the paper it seems amazing
Then Larry Ellison announced two new appliances:
- M6 Big Memory Machine: in that box, 32 TB of DRAM memory is available. It has been designed to work with the new In-memory option and can be linked to an Exadata machine. And it's available right now!
- Database Backup, Logging, Recovery appliance: under that long name hides a new engineered system designed to take care about backups. Only changes are sent to the appliance so the network bandwith needed should be reasonable enough to install the appliance on another site. As an option, the application can synchronize with another appliance on DR site or with the Oracle public Cloud.
The day ended with the welcome reception, directly on Howard Street, which is partly taken over by Oracle for this week. It was the time to eat, listen a music band and relax with a drink.
I hope the next days will be as interesting.