Hummingbird ETL: Genio Is Out of the Bottle with a New Name.
By Michael Reed
This product evaluation concerns Hummingbird Ltd.'s Hummingbird ETL™ v. 5.0.3 product, formerly known as Genio. The product is a valuable tool which has many applications, and has undergone a name change in the process of a re-branding effort. Given Hummingbird's recent strong emphasis on document & knowledge management, our first question was, "Is Hummingbird ETL as a point solution for complex extract/transform/load requirements still a corporate priority as an on-going product offering?" The answer was a solid "Yes". All Hummingbird has to do now is make sure that they raise the level of awareness in North America to match their commitment to the product, and let Genio (now Hummingbird ETL) "out of the bottle".
In a prior life as Research Director at an industry analysis firm, I test-drove and reviewed Genio 2.0 in 2000 and found it to be a robust product with a surprisingly low level of awareness in the North American market. With the Hummingbird ETL 5.0.3 product, I have found both observations remain true.
ETL tools have become something of a commodity in today's market, and are not nearly the niche specialty they used to be. This tempts some companies into thinking that selecting one is a fairly trivial task. This is, however, far from the truth. I spent most of the 1990's as an ETL tool consultant, starting at Trinzic Inc. in 1993. In the early days of ETL tools, the ideal of being able to "move anything from anywhere to anywhere else" was nothing more than a pipe dream. The original ETL tools (Trinzic, Carleton, Prism, etc.) were a far cry from that ideal. Over the years vendors came much closer to being able to provide real business value by accomplishing tasks such as:
- Meta data awareness (early products were "meta data ignorant". They could not look at database catalogs and determine what a table looked like; the programmer had to define the interface by hand. This was a key failure of early ETL tools).
- Propagate only changed records from a database (previously, the programmer had to examine each record and compare to see if it had changed. Over hundreds of thousands of records, this became problematic).
- Examine complex source systems such as SAP R/3. (Previously impossible).
- Move very large datasets quickly (so-called "bulk loading". In earlier years, it was often impossible to move all the data required in the time window available).
- Provide a graphical programming environment (the early tools required custom programming in some type of proprietary language).
- "Proxy" data movement. The capability to move data from source to target without it having to pass through the ETL server machine physically. (Also previously impossible, which meant you had to dedicate a very expensive, server-class machine to the ETL tool).
Over the last five years, in addition to major feature enhancements, many ETL vendors have disappeared. All of the pioneers of the ETL world are in fact long gone. The list of vendors and products which dominate the majority of the ETL market today did not exist even seven years ago.
The Hummingbird Offerings
So, how does Hummingbird fit into this new environment, and where does Hummingbird ETL fall in the product mix?
Hummingbird ETL Feature & Function
Hummingbird's ETL offering is a very robust, highly functional product with an intuitive, fairly easy-to-use graphical interface. In addition to the standard ETL tool features, it has strong capabilities in the areas of meta data integration, mainframe data access, SAP R/3™ data access, and Teradata database fast loads. Hummingbird may not widely advertise the strength of some of their ETL capabilities, but they exist nonetheless. As an example, Hummingbird ETL supports SAP IDOCs. If you aren't an SAP programmer, this may seem like a "gee, that's nice" kind of feature. But in reality, IDOCs allow you to access changed records in an SAP instance without having to examine the entire set of tables involved in the transaction. Just the changes are presented to your ETL interface and can be processed accordingly. This is a very powerful feature, and not one to be ignored. (Ask anybody who tried to work with SAP before ETL tools had IDOC capabilities!) In addition, Hummingbird ETL has achieved SAP certification in every area for which SAP has a certification (you can not certify direct extraction from SAP via ABAP code).
In the area of high speed data loads, Hummingbird ETL is well qualified. As data warehouses have grown larger and extracted data volumes have multiplied exponentially, speed has become a determining factor in product selection. Hummingbird could see this coming, and has been working towards this need since I first looked at the product.
Other features of note in the product include the ability to change program variables during execution, mainframe connectivity via the Striva DETAIL™ product, meta data integration via the Hummingbird ETL™ Met@Data solution, and the ability to bypass the ETL engine and move data directly from source to target. The product also handles database cursors well, and allows user-defined transaction control, including auto-commit. Some of the wizards included with the product are also quite sophisticated. This is not a complete technical evaluation of the Hummingbird ETL product (for more in-depth details, consult Enterprise Warehousing Solutions'™ ETL evaluation white papers), so we will keep the technical list somewhat brief. But we found no significant gaps in the technical offering, and Hummingbird ETL holds up well against all of its competitors.
Hummingbird's Corporate Philosophy
Some may find it curious that Hummingbird ETL is not better known in the North America. You might ask "If the product is so good, why don't I know anybody who's using it?" The answer is that the Hummingbird ETL product's roots are in Europe, where Hummingbird acquired the founder of the product, a French company called Leonard's Logic. They started there, have a strong presence there, (R&D actually resides in Toronto), and their largest customers are also European companies. The company would like to share this secret with the rest of the world.
You have to listen carefully to hear the ETL voice among the other products that Hummingbird offers. Over the last few years, Hummingbird has made major strides into the Document & Knowledge Management arenas, purchasing companies such as PC DOCS and Fulcrum. As stated by the Gartner Group in September 2002, Hummingbird "has moved to create a coherent brand and product structure encompassing content, document management, collaboration, and information access." If a prospective customer were to look at Hummingbird, strictly from the information on their press releases and the general industry buzz, they might think that an ETL tool is no longer at the forefront of Hummingbird's corporate strategy. But if you consider the requirements for true "document management", the ability to extract, transform, and load unstructured data (i.e. Microsoft Word documents, Adobe PDF files, etc.) is an imperative. Knowing this, Hummingbird added capabilities for unstructured data into Hummingbird ETL 5.0, and enabled it as an integral part of the Hummingbird Enterprise™ suite. These changes do not in any way reduce the ETL product's strengths as a point product however. As Hummingbird's management pointed out, "most of Hummingbird's ETL sales are pure ETL sales, a much smaller percentage are currently as a part of Hummingbird Enterprise". According to Mathias Evin, Product Manager for Data Integration, "an integrated solution can be any combination of products… Hummingbird is committed to ETL as a point solution, as well as part of a suite". Examined in this light, Hummingbird ETL actually provides good traction in their Document Management offering, without any change to the product's original function.
Looking forward, the future of ETL tools will be largely based on how well they integrate and facilitate work with complex meta data and XML. Unstructured documents are the latest frontier in data access and integration, since structured document access (i.e. database tables) has become commonplace, and indeed, a commodity purchase. Now that databases come with ETL capabilities (Microsoft SQL Server and its Data Transformation Services), ETL tools come with meta data repositories (Informatica, Hummingbird, and others), and vendors continue to be acquired (Sagent's recent purchase by Group One) - the key to ETL success will be how well the product can bring structure to the unstructured.
One last area to be examined is the strength of the company offering the tool. When a company purchases a software product, they hope the vendor will be around to support them in years to come. As mentioned earlier, many vendors have disappeared from the scene. In some cases their tools were taken over and supported by the acquiring vendor, and in some cases the tools were "stabilized" (meaning no new releases and an eventual loss of support for the product). In Hummingbird's case, they are committed to their product line, have diversified lines of business (which spreads risk) between Connectivity (41% of business) and the Enterprise products (59%), and have a solid cash position. Their stock price has been tracking above the NASDAQ index, and the company has just completed yet another profitable quarter. In the world of software vendors, Hummingbird is as stable as you can hope for.
Hummingbird's ETL product is a strong offering. It has capabilities in tune with the current needs of the market (unstructured data, fast data movement, SAP R/3 access, meta data awareness), commitment from management to raise the level of customer awareness about the product, and more. One thing that impressed us is that Hummingbird is keenly aware of their need to turn up the volume on the marketing of the ETL product, so that North America can get the message Europe has already heard. To us, that indicates a level of commitment that mere words can not convey.