Data Analytics

Think fast with IBM Streaming Analytics

Share this post:

Learn how IBM clients are moving from batch analytics to real-time streams at Think 2018

  • What are some real-world use cases for stream processing technologies?
  • Which companies are using IBM® Streaming Analytics on an enterprise scale?
  • How can I find out more about the potential of stream processing?

Stream processing offers huge potential benefits for organizations that need to respond to threats, opportunities or customer behavior in real time.

Unlike a traditional approach to analytical data processing, where data is collected into a central repository and analyzed in batches, a stream processing engine can analyze each incoming data-point or transaction in-flight, before it even hits the database.

This makes it possible to route data to different systems depending on its content, or to trigger alerts and events almost instantaneously—enabling the organization to become much more responsive to a rapidly changing environment.

Putting it into practice

Stream processing can be an extremely powerful technology, but businesses often struggle to understand when and how to deploy it. For some use cases, streaming analytics may be overkill—if true real-time analytics isn’t really necessary, a simpler batch processing architecture may be sufficient.

However, as competition continues to intensify, we’re seeing an increasing number of situations where streaming analytics can provide a genuine differentiator. In many cases, the first mover in a given market can seize significant competitive advantage by building systems that can respond to emerging situations more quickly than its rivals.

To help companies gain a clearer view of how stream processing can benefit them, we’ve invited a number of industry-leading organizations to share their experiences at IBM Think in Las Vegas (March 19th – 22nd, 2018).

The conference will provide a unique opportunity to listen to some of the biggest names in business walk through their most compelling use cases, and explain how they have used IBM Streaming Analytics to realize their visions.

Real-time customer care

For example, Accenture will be speaking about a new offering called the Customer Care Cognitive X-Ray, which analyzes unstructured data from customer interactions (such as chat logs, email and social media posts) and presents it to customer care teams as intuitive real-time visualizations.

As the data flows into the Cognitive X-Ray system, IBM Streaming Analytics cleans and normalizes it, and then passes it through a set of IBM Watson services. The Watson cognitive technology identifies the relevant entities (products, companies, people, dates, etc.) mentioned in the text, as well as assessing the customer’s intent (what they want to do) and sentiment (how they feel about it).

As a result, the Cognitive X-Ray can be trained to help customer care teams solve a wide range of problems—for example, targeted marketing via customer micro-segmentation, churn prevention, and proactive care based on customer lifetime value. By harnessing IBM Streaming Analytics to provide near-instant insight, the platform enables Accenture’s clients to intervene as soon as a customer expresses a need or raises a complaint—boosting satisfaction, winning trust and reinforcing loyalty.

At the cutting edge of threat analytics

We’re also looking forward to a session from AT&T’s security team, which will describe how the company has integrated Streaming Analytics into its Threat Analytics Platform.

The security team is using stream processing to analyze domain name server (DNS) traffic, and accelerate the detection of compromised servers that may be participating in botnets, exfiltrating data, or engaging in other malicious activities.

AT&T is using Streaming Analytics to move the initial collection and analysis of the DNS data closer to the edge of its network. Instead of waiting until the data has been stored in its central data warehouse, the team is now able to analyze it on-the-fly.

Streaming Analytics passes each of the millions of incoming DNS events into a trained neural network that evaluates whether the event originates from a server that has been compromised. It also uses time-series analysis to monitor the volume of traffic and detect anomalies, such as sudden peaks in traffic that might indicate nefarious activity.

The analysis is completed almost instantly, allowing the system to decide on the best next step. If the data seems normal, it is simply loaded into AT&T’s central data warehouse platforms for standard reporting and analysis. But if it seems suspicious, the threat analytics platform can immediately raise alerts in the company’s operations centers to allow security teams to investigate and implement countermeasures quickly.

Challenge yourself to Think

To learn more about how Accenture, AT&T and other IBM clients are transforming data into instant insight with stream processing, join us at Think 2018 in Las Vegas from March 19th to 22nd—you can register here. Or if you can’t wait to check out the latest features of IBM Streaming Analytics, why not sign up for free access to the latest version.

IBM Cloud Marketing Manager, Global

More Data Analytics stories
November 15, 2018

Watson Knowledge Catalog is Now Available in the Tokyo Data Center

Watson Knowledge Catalog is now available in the Tokyo data center, so customers with requirements to keep their data close to home can now take advantage of servers in Tokyo.

Continue reading

October 30, 2018

IBM Streaming Analytics: Announcements and Deprecations

The IBM Streaming Analytics team is excited to announce additional plans for IBM Streaming Analytics in the United Kingdom and the deprecation of VM-based plans in the United Kingdom and United States.

Continue reading

October 16, 2018

Growing Our Family of Elastic Data Warehouse Offerings

Meet Flex, the newest member of our elastic data warehouse family. It delivers independent scaling of storage and compute, self-service backup and restore, and fast-recovery HA in a configuration optimized for storage-dense workloads.

Continue reading