On June 30, 2006, Luke Liang, CEO and co-founder of InetSoft, was interviewed by Steven Haines from InformIT, about the company's roots in Java reporting tools.
I love meeting with executives that are also strongly technically inclined, which made my interview with InetSoft's CEO and co-founder, Luke Liang, so much fun. In 1996, Luke Liang and Larry Liang created InetSoft Technology, just as Java was coming into popularity. As an emerging market, Java lacked tools, and as they surveyed the market they decided to address the most urgent need: reporting. With hope and engineering prowess, they built reporting tools for extracting data from corporate data sources and presenting it in a meaningful way. And their efforts paid off as they experienced a good reception from the community.
Rooted in a reporting foundation, InetSoft Technology continued to grow and soon became a leading provider of reporting software, even receiving Java Developer Journal (JDJ) awards five years in a row.
As the market evolved, something happened that always happens: whenever you give users a new capability, they want more. Specifically, once executives learn about what type of information is available to them, they want it and a whole lot more. One limitation InetSoft faced is that their reporting engine was very easy to use for the technically savvy user, but it was never targeted at the high-level executive who may not possess the technical skills to build complicated reports. Furthermore, they started gaining popularity in the Business Intelligence (BI) communities where users are interested in the results, not the process required to get to those results. Information Week hosted a survey asking BI professionals about the challenges they face in their industry, and the following are the top three responses:
1. Ease of use
2. Integration of multiple platforms
3.
Data quality issues
InetSoft analyzed these responses and derived the following conclusions:
• Ease of use: they asked themselves if the problem they were solving was so complex that they could not make it easier? The complexity in such a solution existed because the product was targeted at highly skilled technical people, not the business person. As such the business professional needs a new paradigm that is easier to understand in his relative domain.
• Integration of multiple platforms: the data is drawn from different data sources (CRM systems, marketing databases, financial databases, and so forth) and potentially running on different platforms. Therefore, the business professional needs access to be able to manipulate all of this information in a single interface.
•
Data quality issues: the same piece of data may appear in multiple databases, but it could appear with a
different name. The business professional needs a mechanism for tying seemingly disparate data from different
data sources together, given knowledge of the business domain.
Their goal then was to integrate business process intelligence into their front-end interface instead of focusing on one database or one data warehouse. They therefore grew out of the reporting space and into the business intelligence and business process space. And to do this they needed to address three things:
• We will always have different data source to deal with, either through different off-the-shelf software purchases (Oracle Financials, Siebel, etc.), corporate acquisitions, or new in-home applications.
• Databases will be designed on different levels, based off of functionality. For example, operational databases will focus on transactions while data warehouses will focus on high-level aggregate views.
•
Data will be presented in different data types. Some will be common database data types while others will be
XML data and SOAP messages. In order to manipulate this data together, the data needs to be formed into a
common format.
This analysis yielded their version 8.0 product that boasted a new technical paradigm: “Data
Block” technology. This technology is presented in two pieces: a middleware data federation engine that
combines all of the data from disparate systems, and a front-end data worksheet for dynamically assembling
this data in real-time. A data worksheet manipulates data blocks, which can be extracted from different data
sources through whatever means required, such as through SQL queries. Furthermore, the data worksheet uses a
spreadsheet paradigm to allow the user to manipulate this seemingly unrelated data: the data blocks can be
joined, intersected, unioned, and so forth. The point is that the spreadsheet paradigm can manipulate the data
much further than simple SQL queries and involves data from multiple data sources.
This allows users to think about the problem in a different way, a way that focuses on the business need and
not the technical process of gathering the data. Contrasting this to existing processes, consider data stored
in a data warehouse: no end-user hand will be granted access directly to the data warehouse data. Instead the
business owner defines the requirements, such as analyzing strategic goals and measuring profitability. But
based off of the results of those queries, or the sheer dynamic nature of the data itself, the business owner
changes requirements on the IT staff for generating these reports. The result is that everyone is frustrated:
the IT staff is frustrated because the business owner does not seem to know what he really wants, whereas the
business owner is frustrated because it is so much effort to get at the data he wants. Furthermore, when he
gets it, it is dated. The “Data Block” paradigm addresses this in a couple very important ways:
• It provides the basic building blocks of data to the business owners for their own manipulation, in a familiar interface.
•
The data federation does not move data, but rather creates a view into live data.
Because of this, the business owner can use an interface that he is already comfortable with to manipulate
data blocks that reflect live data. Therefore, if they see something of interest, they can instantly dive
deeper to get more information. This makes business owners more productive without burdening the IT staff. Now
the technical skills required to obtain information is no more difficult than using Microsoft Word: it is
dashboard based where data can be selected and "sliced and diced" by simply choosing a couple
fields.
Where are we headed?
As with everyone that I interviewed at JavaOne 2006, I asked Luke where he sees the data reporting industry
moving in the next three to five years and here are his thoughts:
1. The industry will continue to struggle on the ease-of-use front. InetSoft is expending great effort enhancing the end user experience, but as systems become increasingly complex and business processes expand, the paradigm will continue to refine itself to best satisfy the business.
2.
The technology to analyze and search huge quantities of data will evolve. Google's solution to searching data
unfortunately applies only to unstructured data, but reporting interacts with structured data. In order to
successfully
maneuver such volumes of structured data, new technologies will appear.
3. New communities will evolve beyond the context of infrastructure. In order for reporting vendors to
address the vast needs of its customers, standards need to evolve and communities need to be built around
those standards.
Summary
Enterprise reporting is a non-trivial problem that spans far more than application reporting. InetSoft Technology is a forerunner in this market and has altered the playing field through the invention of its “Data Block” technology. The core concept is to move views of live data as blocks from multiple data sources into a spreadsheet paradigm where they can be manipulated by a business owner. The result is greater productivity on all fronts..