Center for Stream Computing

Sponsored by

What is Stream Computing and Streaming Analytics?

Real-time data analytics is becoming a more important tool for businesses seeking to maximize productivity and establish themselves as leaders in the ever-changing landscape of evolving business practices today.

But it becomes obvious that these technologies must evolve to accommodate the specific needs of real-time data processing. This evolution has given us new terminologies — stream computing and stream analytics — that define the importance of accessing data streams in real time and delivering insight to enterprise executives.

Looking at Real-Time Stream Processing

Analyzing streaming data is not a new concept. After all, flight computers, missile targeting systems, medical devices and other critical infrastructure operations require that real-time streaming data be processed quickly and correlated into something that may drive decisions. In many cases, those decisions are executed by algorithms under the watchful eye of a computer, or the data is correlated and presented in real time to some form of display, allowing a human to monitor a process and take action as needed.

However, bringing real-time capabilities to businesses has been a much different story. To have analytics drive value, active real-time data streams must be correlated from multiple sources to identify actionable information, based upon identified patterns and data relationships that have been scoured from those various sources of data.

Further complicating real-time analytics are the large amounts of data being generated today and how that data is moving faster than human interpretation, requiring that some sort of machine learning and artificial intelligence be instituted into the process.

Streaming data can come from a multitude of sources, including industrial equipment, customer-facing web pages, online stores, IoT (Internet of things) -connected devices, sensors, inventory systems, production line systems, financial data, and so on. Businesses today are surrounded by real-time data sources and are awash in streaming data. However, putting that data to use has been somewhat challenging.

Many businesses have invested in business analytics packages and processing systems to realize additional value from collected data, with reasonable success. However, the insights derived and the relationships uncovered have taken the form of static data points. In other words, businesses are looking at large amounts of data to discover what has already happened, and not what is currently occurring.

To garner up-to-the-moment actionable information or drive automated systems, data gathering and processing must be treated differently than the siloed approach of data lakes and big data analytics.

Eliminating the pain points of irrelevant data requires the adoption of new technologies, including the benefits of the symbiotic relationship between stream computing and stream analytics.

Working together, stream computing and stream analytics tackle the velocity of the data, as well as its volume, variety, and veracity, to deliver actionable insights and bring value into the analytics picture.

What Are Data Streams?

Data streams — also known as streaming data — describe data that is generated continuously from numerous data sources and sent as data records simultaneously from multiple sources. Typically it consists of small bits of information, measured in kilobytes. Streaming data comes from a wide variety of sources and can include data elements such as log files, e-commerce purchases, in-game player activity, social network information, financial transactions, geospatial services, as well as telemetry from connected devices, or from instrumentation in production lines, shipping systems, and data centers.

Data streams are processed sequentially via a batch processing system and incrementally on a record-by-record basis or over sliding time windows. They are used for a wide variety of analytics including correlations, aggregations, filtering, and sampling. Information derived from data streams gives businesses visibility into many operational aspects, including customer activity.

The derived information can give insight into service usage, server activity, website clicks, and geo-location of devices, people, and physical goods. That insight enables businesses to respond promptly to emerging situations. For example, businesses can track changes in public sentiment on their brands and products by continuously analyzing social media streams, and respond in a timely fashion as the need arises.

Streaming data processing proves most beneficial in scenarios where new, dynamic data is generated on a continual basis. Most businesses encounter these scenarios in one form or another, and is evident in any business that has turned to big data analytics as a way to realize additional value from stored data. Businesses looking to leverage real-time data analytics usually begin with simple applications, such as collecting system logs, and rudimentary data processing like inventory flow, or shipping logs.

What Is Stream Computing?

In its most basic form, stream computing is identified as a high-performance computing system that analyzes multiple data streams from multiple sources live. However, there is much more to stream computing than this.

Today, stream computing is all about pulling in streams of data, processing that data and streaming it back out as a single flow. Stream computing uses software algorithms to analyze data in real time, while it streams, to increase speed and accuracy when dealing with data handling and analysis.

The thinking behind stream computing emanates from the concepts of big data analytics, where large amounts of data are processed from multiple sources to identify relationships and deliver insights based upon the data. However, stream computing delivers on the idea of real-time data analytics, while big data has been focused on looking at static, stored and siloed data.

Stream computing introduces the capability to analyze data that is in motion, allowing businesses to achieve a real-time response or action based on analyzing the data. Stream computing can handle continuous or “bursty” streams of data, which may amount to millions of events per second with microsecond latency. Stream computing is designed to process any type of data, including structured or unstructured data — such as audio, video, network logs, sensors, and social media feeds. From the outset, stream computing is designed to scale, allowing it to process any size of data, from terabytes to zettabytes per day.

Stream computing changes the whole dynamic of big data analytics by changing the location, timing and quantity of data a business can analyze. Stream computing also introduces agility into the analytics process, allowing businesses to store less, analyze more, and make better, quicker decisions. Stream computing can also potentially reduce costs by analyzing stream data and storing only what is necessary.

Simply put, stream computing provides businesses with the ability to detect and make real-time decisions based upon active data.

What Is Stream Analytics?

Stream analytics, often called streaming analytics, is a more generic way to define what stream computing can accomplish. However, streaming analytics is more about the actual analysis and value proposition offered by real-time data analytics, as opposed to the physical processing of the data. Some analysts consider stream computing and stream analytics to be interchangeable definitions that amount to classifying real-time data analytics.

In its most basic form, streaming analytics can be defined as the ability for businesses to set up real-time analytics calculations on data streams from IoT devices, applications, social media, sensors, devices, websites, and much more. Streaming analytics provides quick and appropriate time-sensitive processing along with user-defined algorithms, machine learning, and policy-based event execution to deliver insight or take action in real time to meet the needs of the business.

Streaming analytics is essentially all about extracting business value from data in motion in the same way traditional analytics tools make use of data at rest.

Advantages of Stream Computing and Stream Analytics

  • Provides deeper insight through data visualization. Visualization makes it simpler for companies to manage key performance indicators (KPIs). KPI data viewed in real time produces a dynamic view of the company’s performance at any given time. That data can be used to improve sales, reduce costs, identify errors, and provide information to react faster to risks. Streaming analytics accelerates decision-making and provides access to business metrics and reporting.

  • Offers insight into customer behavior. Streaming analytics gives companies visibility into what customers’ preferences are and what they are buying or not buying. The company has the ability to better understand customers, maintain relationships and increase revenues through up-selling and cross-selling of goods and services.

  • Remain competitive. Streaming analytics allows businesses to better identify trends, recognize benchmarks, and generate forecasts of the company and industry. That capability reduces internal and external threats and provides awareness of industry changes, leading to innovations that help the company to remain competitive, and strengthen the brand.

Streaming Analytics Use Cases

Stream computing and streaming analytics are designed to deliver on certain discrete business operational needs. Typically they feature the following capabilities:

  • Perform advanced real-time analytics on data in motion via data streams

  • Rapidly ingest, correlate and continuously analyze a massive volume and variety of structured and unstructured streaming data as it arrives from thousands of sources

  • Make real-time predictions and discoveries as data arrives

  • Visualize data easily with drag-and-drop development tools

  • Detect and respond to critical events immediately

  • Learn and update models for future analysis and trend prediction with cognitive computing

Applying those capabilities to real-world scenarios creates an interesting cross section of benefits and capabilities across a variety of industries:

  • Industrial sector: Sensors placed in transportation vehicles, industrial equipment, and farm machinery can send relevant data to a streaming application. That application monitors performance, detects any potential defects, and places a spare part order automatically, preventing equipment downtime.

  • Financial services: A financial institution tracks changes in the stock market in real time, computes value-at-risk, and automatically rebalances portfolios based on stock price movements.

  • Real estate: A real-estate website tracks a subset of data from consumers’ mobile devices and makes real-time recommendations of properties to visit based on their geo-location.

  • Utilities: A solar power company has to maintain power throughput for its customers, or pay penalties. A streaming data application can monitor all panels in the field, and schedule service in real time, thereby minimizing the periods of low throughput from each panel and the associated penalty payouts.

  • Media: A media publisher streams billions of clickstream records from its online properties, aggregates and enriches the data with demographic information about users, and optimizes content placement on its site, delivering relevancy and better experience to its audience.

  • Entertainment: An online gaming company can collect streaming data about player-game interactions, and then feed the data into its gaming platform. It then analyzes the data in real time, and offers incentives and dynamic experiences to engage its players.

Show less
Show more