5 signs an organization should try cloud-based streaming analytics

Manager of Portfolio Strategy, IBM

Without a doubt, software consumption and delivery models are transforming business. Many software vendors are going with a cloud computing–first delivery model at the request of clients. A recent Goldman Sachs study projected that "spending on cloud computing infrastructure and platforms is expected to grow at a 30 percent compound annual growth rate (CAGR) from 2013 through 2018, compared with a 5 percent growth rate for overall enterprise IT."

A survey of other firms comes to a similar conclusion. A Centaur Partners analysis of software-as-a-service (SaaS), cloud-based business application services revenue forecasted “the market growing from $13.5 billon in 2011 to $32.8 billion in 2016 to attain a 19.5 pecent CAGR." And IDC reports public cloud computing may reach $70 billion in 2015.

Why cloud-based delivery makes sense about what kinds of applications are well suited for cloud-based delivery is important. While this delivery mechanism offers immense benefit—and rightfully, our IT spend is supporting cloud—driving everything to the cloud is an extreme measure, which is never a good approach.

As an example, many clients in finance, government and healthcare have highly sensitive data and very strict regulations and mandates that can rule out some cloud-based options. For others, the type of application may dictate the deployment model. Social media analysis is a good application for cloud-based delivery because lots of fast-moving data is typically involved, and that data is already public. Financial forecasting, on the other hand, involves complex priority algorithms that may work best in a hybrid-cloud deployment model in which data can be kept in house but analytics tools such as the IBM Streams financial toolkit can be used. This solution helps capture data streams as they come in at high velocity.

Five signs an organization can use cloud-based streaming analytics

Consider these five signs that indicate an organization may want to procure a cloud-based streaming analytics service:

  1. An organization has the secret sauce built into a data product or business model, and it wants to continuously update the model based on data streams. In this case, a cloud service that is always connected to data sources is needed. The service can dynamically allocate compute power and resources, and it can automate the update of the data product or business model so it is continuously improving.
  2. An organization has an existing application built for customer relationship management (CRM), and it wants to infuse new insight from data streams such as geospatial positioning. A cloud-based streaming service enables the organization to enhance the application by using its language of choice such as Java or C++ with continuous analysis of data streams.
  3. The data of interest to an organization is typically lost because a repository or place to store the data is not clearly delineated, and that kernel of interesting data can be hard to spot. This scenario is common for security—either cybersecurity or real-world physical security—in which 90 percent of the time the status is green, but in a split second the status can change and be easy to miss. Cloud-based streaming analytics can analyze the data in streams without the need to store the data. This capability means the organization can capitalize on the time value of data.
  4. Machine data is used in an organization, or it aspires to use machine data. In a manufacturing and production scenario, for example, equipment in the field emits lots of signals. Many of these data points can be messy or bad reads or they can be lost because of poor network connectivity. An underground, long-wall mining machine, for example, may lose a signal or sustain damage. However, messy or noisy machine data can be sent to a cloud platform, and streaming analytics can be applied to determine if a corrective course of action is needed.
  5. An organization needs to get started fast with new data science projects involving high-velocity data. Many people don’t have time to build connectors to data streams and their own data streams operators. In this case, a cloud-based service is well suited.

Explore a streaming analytics solution

To experiment with streaming analytics from IBM, explore the Streaming Analytics service on IBM Bluemix. As of 9 October 2015, this streaming analytics service is live. The effort is the result of hundreds of clients participating in an open beta program. Tell us what you think.

In addition, experience the full power of the IBM advanced analytics portfolio, including IBM Streams. And be sure to register for IBM Insight 2015, 25–29 October 2015, in Las Vegas, Nevada.