Connect – a Simba Technologies Blog

TBIO - ROLAP on Teradata - smarter, faster, easier

Teradata recently launched their TBIO (Teradata Business Intelligence Optimizer) product.  Martin Willcox from Teradata has a little YouTube video in which he describes a bit more about the TBIO product.  You can watch it here: http://www.youtube.com/watch?v=_o-qVYRxwM0.  Basically, the Teradata Business Intelligence Optimizer has 3 components:

1. Teradata Schema Workbench

2. Teradata Aggregate Designer

3. Teradata OLAP Connector

Schema Workbench lets you define your OLAP schema within your Teradata database.  Aggregate Designer lets you build the aggregate join indexes so that queries on your OLAP schema run fast.  OLAP connector allows Microsoft Excel Pivot Tables to connect to your Teradata database using the MDX Query Language so you can do ad-hoc query and analysis on your data.  No data movement is required.  Everything is done within your Teradata database.

Could Dell buy Aster Data?

A few days ago, I commented on an article about Oracle's shopping spree and someone responded to me about the possibility of Dell acquiring Aster Data.  Worth a quick read: http://seekingalpha.com/article/223107-winners-and-losers-in-data-warehousing.  Here, Ben Kolada at Seeking Alpha indicates that Aster Data is a premier company in the data warehousing space and could be an acquisition target for Dell.  Interesting though that of the 3 deals that Ben discusses:

1. Microsoft acquiring DATAllegro

2. EMC acquiring Greenplum

3. Teradata acquiring Kickfire

Teradata seems to have paid the least.  Maybe a continuation from the days when Mark Hurd was running the company and keeping a tight rein on spending.  Also, Teradata is further ahead in the high end, big data space than Microsoft or EMC so they don't have to make as much of a splash.

Teradata OLAP Connector - MDX Query Language access to Teradata Database

A little over a week ago, I wrote about how Teradata is now supporting direct connectivity from Excel Pivot Tables via the MDX Query Language.  The Teradata product is called Teradata OLAP Connector.  Here is a link to Teradata's brochure for the Teradata OLAP Connector: http://www.teradata.com/t/brochures/Teradata-OLAP-Connector-eb6212/?type=BR.  Teradata is touting the following as advantages to this product:

- real-time data

- reduced IT workload

- continued security

- workload management

- consistent results

- competitive advantage

The reality is that in the past, Teradata has not really supported OLAP.  You needed to have a mid-tier server that sat between the Teradata database and your BI application.  Now, you just build your cubes within Teradata and so there is no mid-tier and no data movement.  Your BI application gets direct access to the power of your Teradata implementation.  Of course, at this time, Teradata OLAP Connector is only supporting Excel Pivot Tables.  However, since this is an MDX query language solution, well, you do the math...

Oracle's shopping spree and who might Oracle buy next...

I was reading a very interesting opinion on sandhill.com by Steven Janisse entitled "Oracle's Mergers and Acquisitions: Who Will Be Next?"  Janisse has a very nice history of the companies that Oracle has acquired since acquiring PeopleSoft in 2004/2005.  He has a very nice graphic that shows the companies acquired and the amount paid by Oracle.  The graphic has a circle for every acquisition with the size of the circle denoting the amount paid.  Janisse then goes on to list potential future acquisitions by Oracle.  Given recent remarks by Oracle President Charles Phillips about Oracle wanting to do more aquisitions, the opinion by Steven Janisse is worth a read.  Janisse mentions companies such as Teradata, Informatica, TIBCO, etc. as potential Oracle targets.  You can read more at: http://sandhill.com/opinion/daily_blog.php?id=31&post=684.

Teradata supports MDX query language

Well, it is now official, Teradata has launched their Teradata OLAP Connector product which allows Microsoft Excel Pivot Tables to connect to a Teradata database.  Basically, you can define a virtual OLAP cube on top of your existing tables within a Teradata database and then use Excel Pivot Tables to do ad-hoc query and analysis on your data.  Excel connects to the Teradata Database using the MDX Query Language.  The really nice part of this solution is that all your data stays in Teradata and you get more options on how to analyze the data.  You get the full power of Excel Pivot Tables and you get the full analytics power of the Teradata database.  Teradata now supports both the SQL and MDX query languages!  You can read more here:

http://www.teradata.com/t/News-Releases/2010/Teradata-OLAP-Connector-Enriches-Direct-Excel-Connectivity-Enhancing-Enterprise-and-Real-time-Analytics/

http://www.simba.com/news/Simba-Builds-Custom-ROLAP-BI-Solution-for-Teradata.htm

http://cwebbbi.spaces.live.com/Blog/cns!7B84B0F2C239489A!7726.entry

Minding your bits...

I note that Window's transition from 32-bit to 64-bit has been an on-going saga for the last few years. The 64-bit variant of Windows XP was a big step forward. Another major milestone was just recently crossed when Microsoft released the first 64-bit variant of Microsoft Office as part of the Office 2010 launch in May.

For those who were around when Windows last made a machine-word-size transition--from 16 to 32-bit--things will be a deja vu of sorts.

I've noticed that this time around, everyone seems to be jumping onboard the biggest, baddest Excel (64-bit 2010) with a lot more enthuasism than I recall the last time around.

One impediment lurking in the background that many people may be caught off-guard by is the availability of the appropriate 64-bit database drivers/providers. Simba's been working hard to catch up here.

So what are people's experience with 64-bit Excel so far? How has your transition to 64-bit been? What sorts of difficulities have you ran into when you made the leap to 64-bit?

The case for a custom ODBC driver

I was talking to a customer recently and they were telling me they have a very high performance database that scales very well, but they have poor reporting.  They were complaining that they spent so much time on building such a great database product and all their customers see is that they cannot get the reports they need.  So what to do?

The answer is simple, build a custom ODBC driver for the database.  The database product already supports SQL so all they need is ODBC.  The customer complained that they didn't have anyone in the company who understands ODBC.  Again, the answer is simple.  I asked one of our engineers to work with the customer and they literally had Crystal Reports talking to their database using the custom ODBC driver in about a week.  The customer was amazed.  They went to one of their users and gave them a demo of the proof of concept ODBC driver.  The user loved it.  Our customer was ecstatic because they now had access to all ODBC compliant applications to show off their database.  Win-win-win.

Windows XP lives on

Interesting that Microsoft has announced that Windows 7 users can "downgrade" to Windows XP for another 10 years.  Seems that enterprises are happy with XP and continue to use it.  You can read more:

http://www.engadget.com/2010/07/12/microsoft-says-74-percent-of-work-pcs-still-use-windows-xp-exte/

http://www.computerworld.com/s/article/9179109/Microsoft_extends_Windows_XP_downgrade_rights_until_2020?taxonomyId=125&pageNumber=1

EMC buys GreenPlum

Interesting news yesterday about EMC's acquisition of data warehouse appliance company GreenPlum.  GreenPlum had of course partnered heavily with Sun Microsystems which was recently acquired by database giant Oracle.  GreenPlum has some interesting customers including eBay, NYSE Euronext, and NASDAQ OMX Group.  GreenPlum plays in the same space as veterans Teradata and Oracle and upcomers like Aster Data, Kognitio, ParAccel, InfoBright, Vertica, etc.  This is a very interesting space which saw Microsoft acquire DataAllegro and coming out with SQL Server 2008 R2 Parallel Edition as well as SAP's pending acquisition of Sybase.  SAP is also working on their own in-memory column-oriented technology and their upcoming HANA appliance.  As the volume of data continues to grow exponentially, the ability to store and analyze large amounts of data continues to be a challenge for every enterprise and so we will continue to see more action in this space.  As the databases scale to be larger and larger, one question that always comes up is how do we access and analyze all of this data.  The SQL query language accessed via ODBC and JDBC drivers continues to be the language of choice in the database world.  The MDX query language accessed via ODBO and XMLA is now the de-facto language of choice for analytics.

I was recently asked who supports MDX and the answer is, everyone.  The major database players - Oracle, IBM, Microsoft, and SAP all support MDX query language.  The major BI players - SAP/Business Objects, Microsoft, Oracle/Hyperion, and IBM/Cognos all support the MDX language.  Microsoft Excel is on virtually every desktop in the enterprise and if you look at Microsoft's recently released Excel 2010, the BI capabilities in Pivot Tables has been extended with the addition of sparklines, slicers, and even write-back.  Excel Pivot Tables talks to analytic databases using the MDX query language.  SQL and MDX are both necessary for analytics today.

More on SAP's Sybase Acquisition

SAP's pending acquisition of Sybase for 6 billion dollars has a lot of people talking.  When I was at Sapphire a couple of weeks ago, this was a topic that a lot of people asked about.  Unfortunately, the standard SAP line was that the deal has not closed so they could not say much more than what was already publicly stated.  This is fair enough since the deal has not yet closed.  However, industry analysts have a lot to say and Stephen Swoyer at TDWI wrote an interesting article entitled "Analysis: Sifting Through the SAP-Sybase Tea Leaves" which I think is worth a read.  There has been a lot written about the mobile aspects of Sybase but Stephen also delves into the data integration side with which there is a fair amount of overlap between SAP BusinessObjects.  You can read the article here: http://tdwi.org/articles/2010/05/26/sap-sybase-analysis.aspx.

Hasso on Hasso - SAP's Hasso Plattner interviews himself

I was at Sapphire Orlando last week and there was a lot of interesting things coming from SAP.  One thing I saw was a video of Hasso Plattner interviewing himself.  He talks about in-memory and performance improvements.  Quite a cute video but worth the 8 minutes to watch it: http://www.sapphirenow.com/video/hasso.html?keepThis=true&TB_iframe=true&height=380&width=500.

SAP acquires Sybase - BI Implications

Boris Evelson had an interesting blog entry about the "BI Implications Of SAP-Sybase Deal".  You can read the blog here: http://www.information-management.com/blogs/business_intelligence_sybase_sap-10017853-1.html.  From my point of view, it is interesting that Boris is pondering about an MDX query language interface for Sybase IQ.  Simba has partnered with SAP to do this for BW and we have partnered with Oracle to do this for OLAP Option.  Simba has also developed a ROLAP technology that allows us to build cube schemas and an appropriate MDX query language interface on top of relational (SQL) data sources.  We haven't tried to do this on Sybase IQ, but it might be interesting.

Oracle Data Warehouse Strategic Update

At the Oracle Collaborate conference last week, I attended a session presented by Ray Roccaforte who is an SVP at Oracle.  The BIWA (Business Intelligence data Warehousing and Analytics) SIG was Ray's idea/initiative.  Ray talked about the Sun Oracle Database Machine, Oracle In-Database Analytics, and Industry-Analytics Solution Sets.  This was a good session and one major point Ray made was that Oracle was bringing the Analytics to the Data rather than the Data to the Analytics.

Ray talked about the Database Machine and some key points were:

- complete data warehouse solution - hardware and software

- extreme scalability and performance

- short time to value

- hybrid columnar compression

- fast hardware

- big pipe

For In-Database Analytics, some key points were:

- bring the algorithms to the data

- Oracle OLAP for multi-dimensional queries and the Simba MDX Provider for Oracle OLAP which provides an open, dimensional interface.  Oracle OLAP is also a summary management solution for SQL based BI applications.  Oracle OLAP is embedded in the Oracle Database instance and storage so it automatically gets the full benefit of the database machine flash cache.

- Data Mining includes a wide range of traditional and cutting edge algorithms developed inside the SQL kernel and Exadata Smart Scan where model scoring is performed in the Exadata Storage Server

Ray's talk was very insightful and another interesting comment he made is that Oracle is positioning to be a solution provider instead of just a component provider.

Russ Tront and I also had a presentation where we talked about the Simba MDX Provider for Oracle OLAP.  If anyone wants the slides from that talk, please feel free to email me.  Russ and Mike van der Velden also did a hands on lab where they had people walk though analyzing their Oracle OLAP cubes using the MDX Provider and Microsoft Excel 2007.  The participants found this session very valuable.  If anyone is interested in a demo of the Oracle OLAP MDX Provider or wants to get a copy of it to try out, again, please feel free to email me.

Oracle Exadata V2 - Sun Oracle Database Machine

We have had a number of customers try out our MDX Provider for Oracle OLAP on the Exadata Database machine and they have said the performance is incredible.  I talked a little about Exadata previously http://blogs.simba.com/simba_technologies_ceo_co/2009/12/oracle-exadata.html.  Last week, I was at the Oracle Collaborate Conference and I took the opportunity to attend a session on the Exadata Database Machine by Oracle SVP Juan Loaiza.  This was a very good session where Loaiza talked about version 2 of this product and its use in data warehousing and OLTP.  A few interesting points from the talk:

1. Loaiza quoted Oracle President, Charles Phillips saying "Exadata is the fastest growing product in Oracle's history" and "The Exadata pipeline is approaching $400 million with Q4 bookings forecast at nearly $100 million."  This is very impressive.  Oracle is a huge organization but even these numbers should start to register on the dial.

2. Loaiza identified 20 Database Machine customers including Coke, RIM, and Thomson Reuters.  Impressive names.

3. Loaiza also talked about Hybrid Columnar Compression.  The "data is grouped by column and the compressed" with two different compression modes - Query Mode and Archival Mode.  Query Mode is for data warehousing with typical compression ratios of 10x.  Archival Mode is for infrequently accessed data with typical compression ratios of 15x and up to 50x for some data.  This compression combined with all the Flash storage and the in storage data intensive processing definitely make this database perform.

Kognitio - row vs column?

I was reading an interesting article by Stephen Swoyer at TDWI where he interviewed the US CEO of Kognitio - John Thompson.  Interesting article that you can read at: http://tdwi.org/Articles/2010/04/21/Kognitio-Row-Based-Databases.aspx.  This continues the arguments of row vs column for analytical databases.  I just got back from the Oracle Collaborate conference and also heard some of the same discussion there.  Oracle now has hybrid columnar compression for their database in version 11g.  One very interesting thing I noted in Stephen's TDWI article is that for Big Data, Kognitio can "do real-time continuous loading into the database".  According to John Thomson, this is "a make-or-break feature" because they have "seen the complete collapse of the batch window".  This is very interesting.  There is no longer a batch window to load huge amounts of data so doing this continuous loading of big data into the database while still being able to perform analytics does sound like a very good feature.

The Kingdom of Excel

Whoa!  At the Gartner BI Summit, they suggested we "Wave the white flag on using Excel for business intelligence".  What breakthrough thinking!  This is something we at Simba have seen for years.  Many BI/OLAP vendors have come to us looking for an MDX Query Language solution for their OLAP/data warehousing products. We have found that a key driver for supporting the MDX Query Language was connectivity with Excel and specifically Excel Pivot Tables.  Microsoft continues to invest heavily in the BI capabilities of Excel and having an MDX Provider for your OLAP/multi-dimensional data source allows Excel to connect natively.  Native Excel connectivity is something we have now brought to Oracle OLAP cubes and we will even be at the Oracle Collaborate 10 conference next week where we have two sessions on showing users how to use Excel to connect to Oracle 11g database: http://www.simba.com/news/Simba-Provides-Free-Training-Connecting-Excel-PivotTables-to-Oracle-OLAP-at-COLLABORATE.htm.

Excel connectivity is a very important feature of any data source be it relational, multi-dimensional, ISAM, object-oriented, XML, etc.  Excel can connect to anything that has an ODBC or an ODBO interface.  Once someone can see how their data can be pulled live into Excel and how an Excel user can then get creative in their tool of choice, it becomes a very powerful reason to have an ODBC driver or ODBO provider.

You can read more about what Gartner is saying here: http://searchbusinessanalytics.techtarget.com/news/2240018042/Gartner-BI-Summit-Wave-the-white-flag-on-using-Excel-for-business-intelligence.

In-memory technologies

In the database warehouse/BI/analytics space, there is a lot of talk about in-memory technology and how it improves performance. In his blog, Boris Evelson explains what specifically is in-memory for various products including SAP BI Accelerator, QlikView, TM1, and MicroStrategy. Worth a quick look: http://www.information-management.com/blogs/in_memory_analysis_query_reporting-10017593-1.html.

How Hardware is Impacting Analytics and Data Warehousing

I was reading an interesting article by Stephen Swoyer @ TDWI entitled "Studies in Scale: Updated Processor Can Take Analytics to New Heights".  Stephen talks about how a lot of companies such as Teradata are using Intel chips in their systems and how the Nehalem processors from Intel are going to impact the solutions from companies like Teradata.  In reality, 64-bit processing and in-memory technologies are really coming into their own as can be seen in solutions from companies like Oracle, QlikTech, SAP, Teradata.  An interesting article that is worth reading.

Interesting analysis about Oracle and Sun with respect to BI and DW

As I am catching up on my reading after the Olympics, I came across an interesting overview of what the Oracle Sun merger could mean to the Business Intelligence and Data Warehousing space.  Stephen Swoyer at TDWI wrote the following article entitled "Analysis: What Oracle-Sun Means for BI and DW" which you can read at http://tdwi.org/articles/2010/02/24/Oracle-Sun-Analysis.aspx.  The idea of the data warehouse appliance is very powerful as can be seen by what Teradata and Netezza have been doing and now with what Oracle is doing.  I am wondering if Oracle will also come out with a BI appliance.  Kind of makes sense that this could be a next step for them.

Top BI Vendors

Now that the Winter Olympics in Vancouver are over, I can focus more on work.  A number of friends and partners visited Vancouver for the Olympics and it was great to see all of you.  Thanks for coming.

I was reading an article in Datamation by Jeff Vance entitled "Business Intelligence Software: Ten Leaders".  You can read it here: http://itmanagement.earthweb.com/entdev/article.php/3867751.  Nice analysis of a number of major BI vendors. The top 5 of course are SAP/BusinessObjects, SAS, Oracle/Hyperion, IBM/Cognos, and Microsoft.  Interesting who Jeff picks as the next 5.  The BI space of course is still one of the hottest and fastest growing areas within the software industry.  At Simba, we have started of 2010 very well with a lot of interest in data connectivity from many BI and analytics software companies.

Tutorial on using Excel Pivot Tables with Oracle OLAP Option

There are a lot of people who currently use the Excel Spreadsheet Add-in for Oracle OLAP 10g cubes. Moving to Oracle OLAP 11g cubes, the spreadsheet add-in has been replaced by the MDX Provider for Oracle OLAP. If you need to learn how to use the MDX Provider for Oracle OLAP with Excel spreadsheets, Oracle has now come up with a nice tutorial that covers "how to use Excel 2007 with Oracle11g OLAP data for easy and powerful ad hoc query and analysis": http://www.oracle.com/technology/obe/11gr2_db_prod/bidw/olap11g_excel/olap_excel.htm#s8.

Oracle and Sun - the new IBM of the 60's?

I was reading an interesting article by Larry Barrett about how Oracle's series of acquisitions over the last 5 years including it's recent acquisition of Sun now makes it equivalent to what IBM was in the 1960's.  An interesting perspective and worth the read: http://itmanagement.earthweb.com/datbus/article.php/3861176/Oracle-Takes-a-Page-From-IBM.htm

Oracle OLAP Newsletter - January 2010

I just got the Oracle OLAP Newsletter for January 2010 and we got a nice write-up on Simba's MDX solution for Oracle OLAP 11g.  The University of Applied Sciences in Wildau, Germany has a Masters course in Data Warehousing.  They use Oracle OLAP in this course and the students used Simba's MDX Provider for Oracle OLAP in their course work. They were able to connect from Excel spreadsheets to their OLAP cubes.  You can read more about it on the Oracle website: http://www.oracle.com/technology/products/bi/olap/olapref/newsletter/oracleolapnewsletter_jan10.html.

Oracle Exadata

Over the last year, we have been doing a fair bit of work with Oracle 11g OLAP Option and the MDX query language.  Although I have not had access to an Exadata machine for testing, people tell me that OLAP runs very well on the database machine and our MDX Provider for Oracle OLAP allows Excel to connect to OLAP on the database machine.  In doing more research on the database machine, I re-read an article by Stephen Swoyer from October this year that I thought might be of interest to others who are curious about the Data Warehousing space and Exadata in particular.  You can find the article at: http://www.tdwi.org/News/display.aspx?ID=9628.  Worth a quick read.  I expect that there will be more Exadata reference customers in 2010.

Happy New Year to everyone and wish you all a happy, healthy, and prosperous 2010.

Also, if anyone is interested, Vancouver and Whistler will be hosting the Winter Olympics from February 12 to 28, 2010.  It will be quite an event.  If you want more information, please visit http://www.vancouver2010.com.  If you are planning on visiting, please give me a heads up.

XMLA for Oracle OLAP 11g

Now that our MDX Provider for Oracle OLAP 11g is released, the team has been working on the next version. The released product supports the MDX query language and the ODBO COM interface.  We now have a proof of concept of an XMLA provider in house.  If anyone is interested in trying this out, give me a shout and we can discuss.