Monday, 25 November 2013

Read-All-About-It: new weekly Oracle Data Warehousing newspaper

Thanks to Brendan Tierney for bringing this excellent online automated news service to my attention….

For a long time I have been wondering how to pull together all the articles from my favourite Twitter feeds, Facebook pages and blogs. Well thanks to Brendan I have discovered a service called Paper.li. This weekend I spent some time setting up feeds from all my favourite sources related to data warehousing, big data. Exadata and other related Oracle technologies. The result is the "#Oracle DW-Big Data Weekly Roundup" which is designed to "keep you up to date on all the weekly sql analytics, data warehousing and big data news from # Oracle". The newspaper is refreshed every Sunday night so that it is ready for Monday morning to read over breakfast. It is the perfect way to start the working week….



Newspaper


if you want to subscribe to this weekly newspaper then go here: http://paper.li/OracleBigData/1384259272 and click on the red SUBSCRIBE link in the top right region of the screen. To give you some guidance on the where all this content is coming from, I am pulling articles from the following sources:
  • Oracle Twitter accounts
    • OracleBigData
    • Oracle Database
    • SQLMaria (Optimizer)
    • CharlieDataMine (Advanced Analytics)
    • NoSQL Database
    • SQL Developer
    • Hardware team
    • BI technology
    • Profit Online Magazine
    • Mark Hornick (R Enterprise)
    • Oracle University
  • Oracle Blogs
    • Data Warehousing
    • Data Mining
    • R
  • Oracle Facebook pages
    • Data Warehousing and Big Data page
Looking for feedback on how useful this is to people as we have so many ways to communicate with you it is good to know what works and what does not work.
If you want to subscribe to Brendan's data mining/analytics newsletter it is here: http://paper.li/brendantierney/1364568794
Now I am off to investigate creating the same thing on Flipboard for all you iPad/iPhone and Android users…..hope to have an update for you on this very soon so stay tuned!

Technorati Tags: , , ,

Get some sun and learn about Pattern Matching at BIWA

BIWA medium

This is your chance to get some winter sun and learn about the new 12c SQL pattern matching! I have just received confirmation from BIWA Summit team that my presentation and hands-on lab on SQL pattern matching have been accepted for the 2014 conference (Jan 14-16 2014 at the Oracle Conference Center, Oracle HQ, Redwood Shores). If you are not familiar with BIWA then here is an overview:

The Oracle Business Intelligence, Warehousing and Analytics Special Interest Group (BIWA SIG) is the leading worldwide association of persons interested in the successful deployment of Oracle Database-centric business intelligence, data warehousing, analytical products, EPM/Essbase and Big Data related features and options. BIWA now has almost 2000 members worldwide. BIWA is a not-for-profit entity affiliated with the Independent Oracle Users Group (IOUG).

BIWA runs all sorts of activities:

  • Webcasts - BIWA holds monthly webcasts presenting BI and DW experts speaking on BI, Data Warehousing and Analytics topics.
  • Conferences - Participates in major Oracle conferences including Oracle Open World and Collaborate. BIWA manages the BI/DW track at Oracle Open World's User Group day.
  • Summit - BIWA's annual Summit enables members to learn from and network with hundreds of other BI, Warehousing and Analytics specialists, a uniquely valuable environment that cannot be duplicated by less-focused conferences. Our coming conference will be the Get Analytical With BIWA Training Days conference-within-a-conference at COLLABORATE 10 IOUG Forum, April 18-22, 2010 in Las Vegas, Nevada. See the BIWA2010 tab above for details.
  • Our second summit, at Oracle World HQ in California in December, 2008, provided over 60 technical sessions, 10 hands-on workshops and 4 keynotes. Over 300 people participated. BIWA's first summit, in 2007, also drew over 300 participants and offered over 40 technical sessions, plus hands-on workshops and keynotes.

If you are a data scientist, data warehouse developer, analyst, data miner, BI reports developer etc then this is definitely the group for you. There is more information about BIWA on their website: http://www.oraclebiwa.org

Back to my sessions…in my presentation I will explore some interesting use cases and provide details of some performance metrics from our own internal tests and from work with customers. The hands-on lab shows how pattern matching using native SQL improves application and development productivity and query efficiency for row sequence analysis. The lab is divided into five sections:

  • Part 1 - Introduction to SQL Pattern Matching
  • Part 2 - Searching for more complex patterns
  • Part 3 - Understanding how SQL Pattern Matching works
  • Part 4 - How to use built-in measures
  • Part 5 - Advanced Topics

This is a great opportunity to get some hands-on experience with the new MATCH_RECOGNIZE clause and the developers who worked on this feature will be available in the room to answer your questions. If you are interested in learning more about our pattern matching feature then make sure you attend this conference.  I am already digging out the sunscreen and looking forward to a week away from the freezing cold British winter and enjoying some warm days of clear blue skies and sun.

At this point you should be saying "Wow, how do I sign up for this conference?" and that is a great question. The details of the 2104 conference, including how to register, are here: http://www.biwasummit.com/ . Hope to see you in San Francisco.

 

Technorati Tags: , , , , , , ,

Friday, 22 November 2013

Using Oracle Exadata to improve crop yields

It is not often you read about how the agricultural industry uses data warehousing so this article in latest edition of Oracle Magazine, with the related video from OOW 2013, on how Land O'Lakes is using Exadata caught my attention:  The Business of Growing, by Marta Bright http://www.oracle.com/technetwork/issue-archive/2013/13-nov/o63lol-2034253.html
A little background on Land O'Lakes: Land O’Lakes is a US company that has grown far beyond its roots as a small cooperative of dairy farmers with forward-thinking ideas about producing and packaging butter. It is a Fortune 500 company and is now the second-largest cooperative in the United States, with annual sales of more than US$14 billion. Over the years, Land O’Lakes has expanded its operations into a variety of subsidiaries, including WinField Solutions (WinField), which provides farmers with a wide variety of crop seeds and crop protection products.
This implementation on our engineered systems highlights one of the key unique features of Oracle: the ability to run a truly mixed operation + data warehouse workload on the same platform, in the same rack. Land O'Lakes uses a lot of Oracle applications which push data into their data warehouse. In many cases we talk about the need for the data warehouse to support small windows of opportunities. For WinField solutions this is exactly why they invested in Oracle.
What makes seed sales unique and challenging is that they are directly tied to seasonal purchasing. “There’s somewhat of a Black Friday in the seed business,” explains Tony Taylor, director of technology services at Land O’Lakes. “WinField is a US$5 billion company that sells all of its seed during about a six-week period of time”. With that sort of compressed sales cycle you need to have unto date information at your finger tips so you make the all the right decisions at the right time. Speed and efficiency are the key factors. If you watch the video Chris Malott, Manger of the DBA team at Land O'Lakes, explains the real benefits that her team and the business teams have gained from moving their systems (operational and data warehouse) on to our engineered systems platform.

Land OLakes
Click on the image above to link to the video. If the link does not work then click here: http://www.youtube.com/watch?v=-MDYI7IakR8 
WinField is maximising the use of Oracle's analytical capabilities by incorporating huge volumes of imaging data which it then uses to help farmers make smarter agronomic decisions and ultimately gain higher yields. How many people would expect Oracle's in-database analytics to be driving improved crop yields? Now that is a great use case!
WinField also drills down and across all the information they collect to help identify new opportunities and what in the telco space we would call "churn" - for instance in January they run reports to identify particular farmers who typically purchases seed and products in November but has not yet ordered. This provides WinField the information it needs in order to perform proactive outreach to farmers to find out if they simply haven’t had time to place an order.
“We’re able to help farmers and our co-op members, even in cases where we’re not sure whether it’s going to directly benefit Land O’Lakes or WinField. Because this is truly a cooperative system, these are the people we work for, and we’re willing to invest in them.” -  Mike Macrie, vice president and CIO at Land O’Lakes. 

For next steps, Land O-Lakes is looking to move to Database 12c and I am sure that will open up even more opportunities for their business users to help farmers. Hopefully, they will be able to make use of new analytical features such as SQL Pattern Matching.
 
Technorati Tags: , , , , , , , , ,

Wednesday, 13 November 2013

ADNOC talks about 50x increase in performance

If you are still wondering about how Exadata can revolutionise your business then I would recommend watching this great video which was recorded at this year's OpenWorld.
First a little background...The Abu Dhabi National Oil Company for Distribution (ADNOC) is an integrated energy company that was founded in 1973. ADNOC Distribution markets and distributes petroleum products and services within the United Arab Emirates and internationally. As one of the largest and most innovative government-owned petroleum companies in the Arab Gulf, ADNOC Distribution is renowned and respected for the exceptional quality and reliability of its products and services.
Its five corporate divisions include more than 200 filling stations (a number that is growing at 8% annually), more than 150 convenience stores, 10 vehicle inspection stations, as well as wholesale and retail sales of bulk fuel, gas, oil, diesel, and lubricants.
ADNOC selected Oracle Exadata Database Machine after extensive research because it provided them with a single platform that can run mixed workloads in a single unified machine:
"We chose Oracle Exadata Database Machine because it.offered a fully integrated and highly engineered system that was ready to deploy. With our infrastructure running all the same technology, we can operate any type of Oracle Database without restrictions and be prepared for business growth," said Ali Abdul Aziz Al-Ali, IT division manager, ADNOC Distribution.
".....we could consolidate our transaction processing and business intelligence onto one platform. Competing solutions are just not capable of doing that." - Awad Ahmed Ali El-Sidiq, Senior Database Administrator, ADNOC Distribution
In this new video Awad Ahmen Ali El Sidddig, Senior DBA at ADNOC, talks about the impact that Exadata has had on his team and the whole business. ADNOC is using our engineered systems to drive and manage all their workloads: from transaction systems to payments system to data warehouse to BI environment. A true Disk-to-Dashboard revolution using Engineered Systems. This engineered approach is delivering 50x improvement in performance with one queries running 100x faster! The IT has even revolutionised some of their data warehouse related processes with the help of Exadata and now jobs that were taking over 4 hours now run in a few minutes.
To watch the video click on the image below which will take you to our Oracle YouTube page:
ADNOC oow
(if the above link does not work, click here: http://www.youtube.com/watch?v=zcRpxc6u5Ic)
Now that queries are running 100x faster and jobs are completing in minutes not hours, what is next for the IT team at ADNOC? Like many of our customers ADNOC is now looking to take advantage of big data to help them better align their business operations with customer behaviour and customer insights. To help deliver this next level of insight the IT team is looking at the new features in Oracle Database 12c such as the new in-memory feature to deliver even more performance gains.
The great news is that Awad Ahmen Ali El Sidddig was awarded DBA of the Year - EMEA within our Data Warehouse Global Leaders programme and you can see the badge for this award pop-up at the start of video. Well done to everyone at ADNOC and thanks for spending the time with us at OOW to create this great video.

Technorati Tags: , , , ,

Wednesday, 6 November 2013

Swiss Re increases data warehouse performance and deploys in record time

Great information on yet another data warehouse deployment on Exadata.

A little background on Swiss Re:

In 2002, Swiss Re established a data warehouse for its client markets and products to gather reinsurance information across all organizational units into an integrated structure. The data warehouse provided the basis for reporting at the group level with drill-down capability to individual contracts, while facilitating application integration and data exchange by using common data standards. Initially focusing on property and casualty reinsurance information only, it now includes life and health reinsurance, insurance, and nonlife insurance information.

Key highlights of the benefits that Swiss Re achieved by using Exadata:

  • Reduced the time to feed the data warehouse and generate data marts by 58%
  • Reduced average runtime by 24% for standard reports
  • comfortably loading two data warehouse refreshes per day with incremental feeds
  • Freed up technical experts by significantly minimizing time spent on tuning activities

Most importantly this was one of the fastest project deployments in Swiss Re's history. They went from installation to production in just four months! What is truly surprising is the that it only took two weeks between power-on to testing the machine with full data volumes! Business teams at Swiss Re are now able to fully exploit up-to-date analytics across property, casualty, life, health insurance, and reinsurance lines to identify successful products.

These points are highlighted in the following quotes from Dr. Stephan Gutzwiller, Head of Data Warehouse Services at Swiss Re: 

"We were operating a complete Oracle stack, including servers, storage area network, operating systems, and databases that was well optimized and delivered very good performance over an extended period of time. When a hardware replacement was scheduled for 2012, Oracle Exadata was a natural choice—and the performance increase was impressive. It enabled us to deliver analytics to our internal customers faster, without hiring more IT staff"

“The high quality data that is readily available with Oracle Exadata gives us the insight and agility we need to cater to client needs. We also can continue re-engineering to keep up with the increasing demand without having to grow the organization. This combination creates excellent business value.”

Our full press release is available here: http://www.oracle.com/us/corporate/customers/customersearch/swiss-re-1-exadata-ss-2050409.html. If you want more information about how Exadata can increase the performance of your data warehouse visit our home page: http://www.oracle.com/us/products/database/exadata-database-machine/overview/index.html

 

Technorati Tags: , , ,

Friday, 1 November 2013

SQL analytical mash-ups deliver real-time WOW! for big data

One of the overlooked capabilities of SQL as an analysis engine, because we all just take it for granted, is that you can mix and match analytical features to create some amazing mash-ups. As we move into the exciting world of big data these mash-ups can really deliver those "wow, I never knew that" moments.
While Java is an incredibly flexible and powerful framework for managing big data there are some significant challenges in using Java and MapReduce to drive your analysis to create these "wow" discoveries. One of these "wow" moments was demonstrated at this year's OpenWorld during Andy Mendelsohn's general keynote session.
Here is the scenario - we are looking for fraudulent activities in our big data stream and in this case we identifying potentially fraudulent activities by looking for specific patterns. We using geospatial tagging of each transaction so we can create a real-time fraud-map for our business users.
OOW PM  2
Where we start to move towards a "wow" moment is to extend this basic use of spatial and pattern matching, as shown in the above dashboard screen, to incorporate spatial analytics within the SQL pattern matching clause. This will allow us to compute the distance between transactions. Apologies for the quality of this screenshot….hopefully below you see where we have extended our SQL pattern matching clause to use location of each transaction and to calculate the distance between each transaction:

OOW PM  4
This allows us to compare the time of the last transaction with the time of the current transaction and see if the distance between the two points is possible given the time frame. Obviously if I buy something in Florida from my favourite bike store (may be a new carbon saddle for my Trek) and then 5 minutes later the system sees my credit card details being used in Arizona there is high probability that this transaction in Arizona is actually fraudulent (I am fast on my Trek but not that fast!) and we can flag this up in real-time on our dashboard:

OOW PM  3
In this post I have used the term "real-time" a couple of times and this is an important point and one of the key reasons why SQL really is the only language to use if you want to analyse  big data. One of the most important questions that comes up in every big data project is: how do we do analysis? Many enlightened customers are now realising that using Java-MapReduce to deliver analysis does not result in "wow" moments. These "wow" moments only come with SQL because it is offers a much richer environment, it is simpler to use and it is faster - which makes it possible to deliver real-time "Wow!". Below is a slide from Andy's session showing the results of a comparison of Java-MapReduce vs. SQL pattern matching to deliver our "wow" moment during our live demo.
OOW PM  1
 You can watch our analytical mash-up "Wow" demo that compares the power of 12c SQL pattern matching + spatial analytics vs. Java-MapReduce  here:
OOW PM  5
You can get more information about SQL Pattern Matching on our SQL Analytics home page on OTN, see here http://www.oracle.com/technetwork/database/bi-datawarehousing/sql-analytics-index-1984365.html. You can get more information about our spatial analytics here: http://www.oracle.com/technetwork/database-options/spatialandgraph/overview/index.html
If you would like to watch the full Database 12c OOW presentation see here: http://medianetwork.oracle.com/video/player/2686974264001

Technorati Tags: , , , , , , , ,