Blogs

The Untapped Opportunity of Streaming Analytics

Forrester Research evaluates the IBM big data streaming analytics offering

Manager of Portfolio Strategy, IBM

Analytics is a hot topic these days. It captures tremendous momentum and enables data-driven decisions. And now, real-time analytics is pervasive. Many organizations are evolving their focus on streaming data sources to garner real-time insight. In fact, according to recent Forrester Research survey data, there has been a 66 percent increase in the use of streaming data over the past two years.1

This shift in focus can be attributed to the need to supplement historical or batch analysis with information about what’s happening right now. For example, how do weather or politics impact stock trades, or how can retailers deliver offers to customers as they walk by the store?

Traditional analytics usually depends on historical data that must be persisted and then analyzed. In-memory databases can speed up results and are designed to minimize complexity through columnar data stores, but a batch or reconciliation process is still required at some point. In addition, queries drive the analysis, and analyzed data types may be limited. As a result, traditional analytics is sometimes referred to as rearview-mirror insights.

Streaming analytics shatters traditional models and transforms organizations through forward-thinking approaches. Both traditional and streaming analytics are important. They complement each other to help organizations understand and act on business opportunities. However, in today’s customer-centric, universal marketplace, opportunities may come at a moment’s notice and require real-time action. The capability to speed ahead is important to help avoid missing such opportunities before they are gone for good. The speed of business is increasing, and real-time streaming analytics helps organizations keep pace.

 

Exploring providers of streaming analytics

In evaluating these developments in the streaming analytics space, Forrester Research delivered the report, “The Forrester Wave™: Big Data Streaming Analytics Platforms, Q3 2014.” IBM was among the organizations that Forrester Research invited to participate in the report. In the evaluation, IBM was placed in the Leaders category, scoring highest on performance and scalability optimization based on its IBM® InfoSphere® Streams analytics platform. IBM was also cited for its comprehensive stream processing operators and development tools that can satisfy the “gnarliest of use cases,” according to the report.”2

Seven vendors were evaluated based on 50 criteria in three categories: current offering, strategy, and market presence. Forrester conducted reference calls, completed hands-on evaluations, and reviewed vendor surveys to assess the strengths and weaknesses of each platform. IBM received the highest possible score in several areas, including the ability to execute, implementation support, and input and output data sources.

A Forrester Wave compares vendors and vendor offerings associated with a particular market, measuring the organizations against a set of criteria defined by Forrester. Vendors that cannot meet a set of core requirements are not evaluated.

The evaluation of IBM was based on deployments of InfoSphere Streams for diverse use cases across different industries for extremely fast performance. InfoSphere Streams provides the following features:

  • Speed: Up to 12.3 times more throughput utilizing up to 14.2 times less resources than open source Apache Storm3
  • Deep analytics: Analytics comprising cognitive computing, machine learning, predictive analytics, and prescriptive analytics
  • Integration: Existing analytics models and data management platforms such as IBM SPSS® predictive analytics software and IBM Cognos® business intelligence software

InfoSphere Streams offers a streaming analytics development platform with both a highly visual and intuitive interface and a scale-out runtime. Operators can be leveraged to filter, aggregate and correlate, and enrich any data type. Operators can also be custom built to leverage data types through the InfoSphere Streams open source project on GitHub.4

InfoSphere Streams provides pre-built analytics accelerators developed by mathematicians and scientists in the IBM Research Lab.5 These analytics accelerators cover a broad range of analytics such as R analysis, image recognition, machine learning, sentiment analysis, and more.

 

Exploiting streaming analytics in real time

According to Forrester, “exploiting perishable insights is a huge, untapped opportunity for firms” and “streaming analytics platforms are ready for action.” Forrester also reports that “open source is hyped, but commercial vendors got the goods.”

Visit the InfoSphere Streams website for more details. And please share any thoughts or questions in the comments.

1The Forrester Wave™: Big Data Streaming Analytics Platforms, Q3 2014,” by Mike Gualtieri and Rowan Curran, Forrester Research, Inc., Doc #113442, July 2014. The increase in use of streaming data over the last two years is based on a Forrester survey of 1,658 respondents in May 2014 and documented in “Business Technographics Global Data and Analytics Survey, 2014,” Forrester Research, Inc.
2 “The Forrester Wave™: Big Data Streaming Analytics Platforms, Q3 2014,” by Mike Gualtieri and Rowan Curran, Forrester Research, Inc., Doc #113442, July 2014.
3Of Streams and Storms: A direct comparison of IBM InfoSphere Streams and Apache Storm in a real-world use case – email processing,” IBM Software Group white paper, June 2014.
4 IBMStreams, GitHub website.
5 IBM Research website.