Archive for the Vendors Category

Data Warehouse Appliances - Apples to Apples (update)

I’ve updated the data warehouse appliance spreadsheet, adding two new solutions (SAND & Calpont).  I’ve also updated information on several existing vendor solutions.

Analytics in the Cloud

I attended a webinar today sponsored by Amazon and Vertica called “Data Analytics in the Cloud“.  The Vertica portion was mostly a duplicate of a prior web cast, but the Amazon portion on the Cloud concept was very interesting.  The key points of the cloud concept as I see it are:

  • Pay as you go model - you only pay for the disk space and processing you consume. No start-up costs, but you have to sign a contract. (They claimed the cost was 1/2 that of an in-house solution)
  • Time to market - hours instead of weeks to turn up a terabyte sized system, including hardware, OS, and the Vertica column based database
  • On-demand scalability - seamlessly scales to meet your demand
  • Proven platform - hosted by Amazon on the same platform that hosts the Amazon.com site

I think the benefits of this approach are obvious, especially for a small but rapidly growing operation.  The infrastructure and software license costs alone would be prohibitive, and time to market is critical especially when launching a new idea.

The downsides include:

  • Security concerns, especially for highly sensitive customer data
  • Performance - both in terms of loading large amounts of data and in real-time queries
  • Long-term cost - as with any usage based cost model, the upfront savings could be surpassed by subsequent usage fees

360DegreeVendor.com Site Launch

I’ve launched another site, this one focused on vendor specific information.  The site will provide information in three areas:

  1. Indices - includes BI25 market index, number of vendors in the space, and venture capital metrics (total Venture Capital investment, number of VCs actively invested, number of vendors with VC backing)
  2. Reports - includes lists of vendors broken out by corporate type, offering categories, and venture backing, list of venture capital firms and their vendor recipients
  3. Search - capability to search by vendor name, venture capital name, and offering name to get a specific subset of data (to be available in the next 30 days)

There is some overlap with 360DegreeIndex, particular with the first category.  But the focus of this site will be on the vendors and the information I’ve captured around each vendor (corporate type, offerings, venture backing).

Please e-mail me with any comments, questions, or additions to the information provided.

360DegreeIndex - Sizing the BI and Data Management Market

I “launched” a new web site today, 360DegreeIndex, that will provide market sizing information for the business intelligence and data management vendor space.  “Launched” meaning I pressed the button in FrontPage to populate the site.  It’s not the most aesthetically pleasing site, but I’ll work on that in subsequent releases.  My initial focus is on getting the data published.

The initial version lists data for the following areas:

  • “BI25″ - index and market capitalization for the largest (by adjusted market cap) publicly listed vendors
  • Vendor count by corporate type (e.g., Listed, Venture Backed, Partnership)
  • Venture capital sizing including the number of VC firms (currently 218) with investments in this space, the number of vendors with venture backing (currently 135), and the total estimated current amount of venture capital invested in this space (currently $1.4B)

I plan on updating this data on a monthly basis, and will include month-over-month changes to these figures to provide a view into the trends, which is really the most important part of this exercise. I’m also developing some additional metrics from the 360DegreeVendor database which I will add to the site.

Data Warehouse Appliances - Apples to Apples

I’ve listened to several appliance web casts in the past 2 weeks, and have increased my knowledge of the growing data warehouse appliance market.  One thing that is clear to me is that the landscape has changed considerably in the past 2+ years, with the addition of several vendors providing a “full stack” similar to Teradata and Netezza.  There are also a number of vendors that are providing “software appliances” (to borrow a phrase from Philip Russom, TDWI analyst), by customizing database, operating system, and integration software with either off-the-shelf hardware components (Intel, AMD) or partnering with hardware vendors such as Sun.  And while I’ve gotten bits and pieces of information on this topic from a number of sources, I haven’t seen a full side by side comparison of these products. 

So I pulled together a data warehouse appliance spreadsheet that compares vendors along several categories, including applications, software, architecture (basically SMP or MPP) and hardware.  I’ve also attempted to group the vendors, using “full stack”, “database”, and “hardware” as my initial cut.  I intend this to be a work in progress, and will provide updates as I add new vendors and refine information for existing vendors.  Once I get a complete set of data, I will provide a summary of the dataset.

Please feel free to send me an e-mail if you see incorrect or missing information, or think I should add a vendor to the mix.

Business Intelligence Market Index - “BI25″

The data collected in my BI vendor database, 360DegreeVendor, has allowed me to create an index that measures the publicly traded business intelligence, data warehouse, and data integration vendor space.  The index captures the adjusted aggregate stock price and market cap for the top 25 BI publicly traded vendors.  I say adjusted because for each vendor, I have determined a factor that represents the percentage of business derived from BI products, services, and research, and have applied that factor to their overall stock price and market cap.

Index as of COB April 29, 2008:

BI25 Index: $301.07

BI25 Market Cap: $47.3B

The numbers by themselves don’t have much meaning, the trending will be the most telling. I will publish these numbers on a weekly basis going forward, and will provide trending reports once I’ve built up sufficient data to make it meaningful.

LogiXML

I met recently with Arman Eshraghi, CEO and Founder of LogiXML and wanted to provide an overview of their company and offering, Logi 9.  LogiXML is a primarily a privately held company (with a recent $5M investment from Updata Partners), with headquarters in McLean, VA.  They list well over 100 current customers, in 13 industry groups. 

Logi9, released in January of this year, is a Java based enterprise reporting platform.  It includes report designer, report distribution, dashboard, and OLAP capabilities, which are also available as standalone products.  The platform also includes a data source mapping and ETL feature that allows for access to a variety of disparate data sources.  Logi9 was built from the ground up on the .Net platform, and utilizes an XML based architecture.  They support standard interface protocols such as SQL, ODBC and JDBC, so they connect to most data source applications (e.g., SQL Server, Oracle, MySql).  Logi9 is server based and leverages standard web technologies such as AJAX – there is no client side code which makes software upgrades easier since you don’t have to worry about pushing code out.  The combination of the XML foundation and the server based architecture allows LogiXML to distribute upgrades and patches to the platform with little or no impact to client applications.

 

They offer a free of charge scaled down version of their product, available at FreeReporting.com.  This is a complete reporting tool aimed at developers and power users.  The support is self service, in that you have to diagnose your own problem via the developer network or discussion forums that LogiXML supports.

 

Their implementation consulting services are primarily focused around installing and integrating the Logi product in the client environment.  They offer both on-site training in McLean and on-line training, with courses in product administration and report building.

 

LogiXML is following a recent trend by forming partnerships with data warehouse appliance vendors Vertica and Dataupia to integrate with their products.  Their partner program includes OEM, reseller, and alliance partners such Vertica and Dataupia.

 

While Logi9 is relatively new (released in the past 3 months), the LogiXML product has been around for almost 6 years, which leads me to believe that the product is stable.  The one concern I have is around the server based architecture – performance on the client side may suffer if the network speed is lacking, since there is nothing stored on the client.  However, I think the trade-off in ease of upgrade is well worth the response time concern, particularly since most networks these days are sufficiently fast to handle this additional data.

Dataupia

I spoke with Jim McManus, VP Channels and Alliances, from Dataupia last week and wanted to share my impressions of the company and primary offering, the Satori Server.  Dataupia is based in Cambridge, MA, and has received venture funding from Polais Venture Partners, Fairhaven Capital Partners, Valhalla Partners, including a recent $16M infusion last fall.  The company was founded in 2005, and went general availability with the Satori Server in May, 2007.  Foster Hinshaw, the CEO, is known as the “Father of the Data Warehouse Appliance”, and founded Netezza before leaving to start Dataupia.  They list several customer success stories on their web site, and currently have 4 customers according to Jim.  They recently released a press release with Subex announcing the deployment of a 150TB Oracle OSS system on Satori Server.

The Satori Server is their flagship product, and is touted as a true data warehouse appliance (think Netezza and Teradata).  It includes server, storage, and “optimization software” bundled into one package.  The hardware is built from common off the shelf components, but is engineered to handle data intensive operations. The “optimization software” includes a scaled back Postgress open source relational database management system, and a vanilla Red Hat Linux operating system.  The system is configured in a shared-nothing or MPP architecture.  They currently store data in a row alignment, but are considering offering column based storage.  According the Jim, the system is currently targeted for operational reporting purposes, which lend itself to row based processing.  Satori Server performance enhancement relies heavily on a dynamic aggregation capability that is defined at loading time.

The differentiator for Dataupia is their “Omniversal Transparency” – basically they support queries in Oracle, DB2, and SQL Server natively.  As an example, take a data warehouse sitting on an Oracle database.  The data warehouse might have a large number of tables, most of them reference or smaller volume tables along side several large “fact tables”.  Once the Satori Server is connected to the network, the Oracle DBA can map the larger volume table(s) and migrate all data.  After the migration is complete, queries run as normal within Oracle, with processing on the larger tables handled within the Satori Server.  The benefits of this approach are significant, as it requires no change to applications currently built to access the existing RDBMS.  The DBA can handle most of the application monitoring from the existing RDBMS console, although Dataupia does include a console for configuring and monitoring the Satori Server directly.

Dataupia provides a number of benchmark results on their web site, and include load times (70 MB/sec), refresh rate, and drill down capabilities (24 months of detail data accessed in 5 sec).  They have not as yet released TPC-H results (planned for this summer) which allow for a more direct comparison to their competitors in this space.  A potential concern is the reliance on data aggregation to boost performance results, which brings scalability into question.  A heavy reliance on aggregates can result in significantly increased data sizes to support a mixed reporting environment, particularly a heavy ad-hoc user base.  The Subex announcement provides one scalability data point, but I’d like to see some additional production implementations before eliminating this concern.

Dataupia offers consulting services, which consist of a 1 week jumpstart engagement that hooks up the Satori Server, does initial configuration, migrates selected data, and runs a number of tests to ensure everything is working properly.  They rely on their system integrator partners to handle any required work outside the scope of this jumpstart effort.

They have a well defined partnership strategy, with three categories:

Ø     Solution Partners that extend their product in areas such as business intelligence and data mining

Ø     System Integrators that work with customers to implement the Satori Server within the larger data warehouse initiative

Ø     Technology Partners that provide complementary hardware or software components that are integrated into the product

In summary, a customer with an existing legacy system should strongly consider this solution, due to the enormous cost savings associated with not having to migrate data integration, reporting, or other existing applications.  This represents not only a labor cost savings, but an opportunity savings reflected by limiting impact to the existing analytical business processes in place.  The greenfield situation is not as clear, due to the lack of industry benchmarks and the fact that Satori Server is not a stand alone solution.  In either case, scalability needs to be thoroughly tested as this is a new product with a small installed customer base.

Venture Capital activity in Business Intelligence

I pulled the following statistics from my vendor database 360DegreeVendor:

  • There are 218 Venture Firms with active investements in business intelligence, data warehouse, & data integration vendors
  • Average total investment in each vendor is US$10.3M
  • There are 135 vendors with active venture backing
  • $1.4B total estimated active venture funding in the business intelligence, data warehouse, & data integration space
  • 1.8 average vendor investments per venture firm
  • Business Activity Monitoring product vendors had the highest percentage (56%) of venture funding for offering categories containing 10 or more vendors
  • As expected, vendors with a Reporting offering had the highest total number of venture backed vendors (30)

Elixir Technology

I was introduced to Elixir by Brooks Rembert from JNetDirect recently and wanted to share my impressions of the company and their primary offering, Elixir Repertoire 7.  Elixir is a 10 year old privately held company based in the Republic of Singapore.  They claim over 600 customers world-wide, most in the Singapore and the Far East.  They have a very robust partnership program in place, with Channel and OEM partners that resell their product, and service partners that assist clients with implementations.  JNetDirect is the exclusive distribution partner in the United States. 

Their flagship product, Repertoire 7, is a Java based reporting and dashboard tool that has a data integration capability as well.  I was recently given a demo of their product by the CEO, Shih Hor-Lau.  He walked me through the reporting capability, which includes support for writing and distributing standard reports along with an OLAP capability to support power users.  They also have a very flexible dashboard capability, with a number of widgets and gadgets available out of the box.  The product was developed from the ground up with Java, but includes a set of APIs (“REST APIs”) which support integration with Java, .NET, Perl, AJAX, Ruby, Flex.  The data integration component, “Data ETL”, is actually an Enterprise Information Integration (EII) tool that allows source systems to be mapped, and data to be extracted and integrated at report run-time.  They do not currently have an underlying meta-data repository, but Hor-Lau said they are planning to add that in a future release.

Elixir has a small professional services organization that is capable of handling installation and configuration of the product.  They rely on their service partners or the client to handle the end-to-end implementation activities.

JNetDirect is currently handling Level 1 support in the US, and is ramping up to provide Level 2 support in the near future.  All Level 3 support will be handled by Elixir out of their Singapore offices.

In summary, Repertoire seems to be a viable alternative for someone looking for a low-cost integrated reporting and dashboard development tool.  The data integration component provides the ability to map into a variety of data source types, but is not a heavy duty ETL tool and should not be used to handle large data volumes, complex transformations, or data cleansing.  The lack of a meta data repository is a concern, as it limits visibility to what is available in the Elixir tool, and prevents you from integrating other tools such as a third-party ETL or meta-data management (MDM) product.

Reference:

-        Interview with CEO on SYS-CON TV at Java1 2006