Data Modeling Lends a Hand to NoSQL Databases

By on

jz_dmnosql_092916Sectors such as banking, healthcare, and insurance are increasingly interested in NoSQL databases and Data Modeling for NoSQL. As they take their first or their next steps in the space, they want “to have the tools available to them that give credibility and legitimacy to the technology,” says Pascal Desmarets, Founder and CEO at Hackolade, a startup focused on helping businesses feel more secure about using alternatives to legacy relational databases.

What’s the concern? Speaking from personal experience as an IT strategy consultant with roots in the relational database world, Desmarets refers to the issues he faced when moving from designing small applications in MongoDB to designing complex ones in the environment.

“I wanted the rigor to organize myself so that I wouldn’t get in trouble because the data model is at the heart of applications, and if you get that wrong you end up with much higher TCO and have to redo things,” he says.

Working with MongoDB and NoSQL document databases, he confronted the issue that Data Modeling ends up in the code of the application that reads the data (for flexibility and scalability reasons), and so too does the responsibility to avoid data anomalies. This, he says, presented an obstacle in terms of being able to have a dialogue with analysts, DBAs, and developers when it came to evaluating design options upfront and recognizing potential issues before committing significant development efforts.

That led him on a search for NoSQL tools similar to ones he had worked with on SQL projects – solutions that would provide the blueprints and visualization of data descriptions, constraints, and relationships that were so useful to charting and changing applications. After all, even non-relational databases, the company points out on its website, have relationships.

“A picture is worth a thousand words, but I couldn’t find this specifically for NoSQL document databases that use JSON documents to store records,” he says. And as applications grew in complexity, mining data and discovering structure within it didn’t seem intuitive or productive to him. “How do you query a database if you don’t know what fields exist that you can query on?” he asks.

A Look Inside Hackolade

Desmarets says he actually tried to get vendors interested in developing a tool like the one he was looking for, with no success. He became, he says, an accidental startup, taking a chance on developing a prototype and gauging the response to it by those he showed it to. He was happily surprised.

People told him that they’d been looking for just such a solution for years – especially those individuals who were in Data Governance and Architecture roles. “I asked them if they liked to look at code, and they said, ‘No,’” he relates. With that encouragement, he took development further and launched Hackolade at MongoDB World 2016 in June in New York.

One of the key Data Modeling value propositions Hackolade offers compared to traditional entity relationship or modeling tools is the way it balances entity relationship theory and JSON schema to leverage the powerful nature of JSON and its capabilities of infinite nested objects like subdocuments and arrays. The software uses a JSON Schema notation that is easily visualized and maintained in a tree view “in a way that is natural to users,” he says.

“This becomes a map for your Big Data. Then it’s easier for users who aren’t coders to say, ‘Here’s how I can mine the Big Data. I now understand the presence of the fields,’” he says. He’s filed a patent application for the feature.

Aligned with a Changed World of Application Development

Desmarets says that Data Modeling for NoSQL might not be such a big deal in a world of applications that stay pretty much the same over time. But that’s not the world that most industries live in today, including big name companies in major verticals. Leaders in financial services, healthcare and insurance, and other critical domains – not to mention tech startups that themselves have grown to become world-class brands – are dealing with Agile development environments where applications are in a constant state of evolution.

“Their apps change every week – in a DevOps culture you even do multiple releases per day,” he says. “To answer the new needs of customers.” There have to be plans in place for a data model to support the ability to follow those changes, and the discussions that have to take place among business and functional analysts, designers, architects, developers, and DBAs to ensure the work is done right.

“The Data Modeling document is going to be what enables that dialogue to be more efficient, and in the end give the company the advantage of being more nimble, of being able to respond faster to requests and of lowering its TCO,” Desmarets says.

For companies that already have begun migrating to Big Data in NoSQL environments and have the MongoDB app to show for it, Hackolade also offers a reverse engineering function to facilitate the modeling of existing instances. It will do a statistical sampling of the documents in the database, infer the schema from that and import it into the tool.

“Analysts can then enrich the model by adding comments and constraints and business rules,” he says. From then on, the organization has the foundations for the dialogue in place, to make changes interactively, conduct what-if scenarios, and evaluate different options. “Then, settle on a design and click the button to generate code for the developer to copy and paste and include in the app,” he says.

Coming up for Hackolade

Currently, the tool is available for a 30-day free trial download, and Desmarets says he expects to formalize some licensing deals with customers soon.

Among the functions his team is working on adding in the short term is to come up with a generic way to port the application to support NoSQL products that work with JSON documents, such as MongoDB, Amazon Web Services, Oracle, and Microsoft databases. While these companies haven’t formalized any deals with Hackolade yet, “it’s pretty exciting for a startup,” he says. Potential clients in the financial services industry are also interested in the addition of features such as LDAP authentication support.

While he acknowledges that not every company is going to take the NoSQL path, he thinks in most cases that failing to do so is a missed opportunity.

“I view my tool as something to reassure CIOs of major companies that they can get into this technology and get great benefits from it,” he says. “Agile development doesn’t mean chaos and NoSQL Some people need convincing about that though, and a tool like this helps.”

Leave a Reply