This content is part of the Essential Guide: SOA tutorial: Trends, governance and the microservice impact
Manage Learn to apply best practices and optimize your operations.

Why mix data analytics with BPM and SOA?

Virtualization and the cloud have enabled organizations to process staggering amounts of data by unbuckling server-side limitations. Today, companies handle billions of pieces of data per day, intensifying the challenge of analyzing that data. Experts predict organizations that integrate data analytics technologies with business processes and enterprise architectures will gain a competitive advantage in 2016 through the use of statistics-backed user and market information.

“IT leaders and SOA influencers are using data that they own and can acquire as a means to develop new revenue streams,” said Good Data CEO and founder Roman Stanek. “This data is a blend of homegrown, CRM, social, SaaS and IoT. The analytics used to derive monetary value from this blend will become a new digital currency.”

But to bankroll this digital currency, organizations have to overcome the big challenges, according to Rob Terpilowski, software architect at transportation logistics firm, Lynden Inc.

“It is not only the handling of this amount of data that will continue to be an important trend this year, but [also] the analysis of this data and the ability to ask questions of and find patterns in this data,” said Terpilowski.

Data analytics can be very useful in business processes, according to Stanek, who proposes that adopting metrics into BPM can result in better decision making. Determining how to integrate data analytics into BPM, however, isn’t an easy process, and organizations that wish to do this successfully should study the approaches that can help make good use of data analytics and glean meaningful information from data tools.

Data analytics use cases and approaches

Facebook processes 4.75 billion of pieces of shared content per day. What’s the use of analyzing that much data? For one thing, organizations of this size often profit from user information by selling small portions of it to other enterprises. They also analyze this data to improve customer experience.

To exploit the opportunities in data analytics, hiring or using consulting data scientists may be the answer for businesses. Bringing data scientists into the software development lifecycle will make it easier to take advantage of new architectures for optimization and machine learning from vendors like Microsoft, Amazon and IBM, said Staten.

For developers who need to embed data analytics into the applications they build, the R and Python programming languages offer some notable advantages. R is an open-source language tuned for predictive analytics and data visualization. Python is a general purpose language, but has strong functionality for collecting and analyzing data. Another useful language may be Julia, a technical computing language which isn’t used as widely for data analytics as R, but whose proponents say it offers more efficient data processing.

Taking app analytics to the cloud

As vendors clamor for more cloud applications, enterprise architects should develop plans to smooth this transition. Kurt Milne, vice president of product marketing at the cloud management firm CliQr, points to companies “replacing customer managed service tiers by adopting cloud provider services” as a major SOA trend this year.

The use of application-centric automation tools can help with this effort, as it can enable services to be delivered quickly without the DevOps learning curve typically required with traditional methods, Milne said. For example, AWS DynamoDB and RDS are becoming popular methods for businesses to easily provision services to circumvent the wait times associated with manual processes for engaging IT resources.

“These tools give the benefits of fast and automated deployment, and they also facilitate application migration between clouds, which avoids cloud vendor lock-in,” Milne said.

Milne and others noted that cloud-based analytics distribution platforms (CADP) can help manage not only the initial cloud application deployment but also future iterations. CADPs deliver very broad functionality, ranging from collaboration tools to self-service analytics. Many major business intelligence (BI) systems vendors — BMC, IBM, etc. — offer CADPs. Examples of relatively new players include GoodData Analytics Distribution Platform (2007) and BIME (2009).

By George Lawton and Jan Stafford