Big Data Analysis Goes SaaS

10 Wacky Kickstarter Projects That Succeeded

If you're ready to pull insights from big data, but analytics isn't a core competency of your business, what's the right move? A Long Island, N.Y.-based software-as-a-service (SaaS) provider is hoping you'll outsource your analytics chores to its new streaming analytics service rather than tackling the job in house.

A spinoff from IT management software provider Nastel Technologies, jKool unveiled its eponymous big data analytics service at this week's All Things Open tech conference in Raleigh, NC. Accessible at jKoolcloud.com, jKool's service is designed to analyze and uncover hidden patterns in time-series data, which the company defines with "the six w's:" what, where, why, and when something happened (that's four), plus "what else happened" and "what I do."

"We see this as a service," said Charley Rich, jKool's vice president of product management, in a phone interview with InformationWeek. "You might be building applications, and instead of building in the analytical functions, you'd call out to us as a service, get the analysis, and spend your dollars on building applications, not on building analytics."

The jKool service is free to try during its soft-launch period. "Our preview pricing is free for up to a billion data points for 14 days," said Rich. "What's a data point? It might be a cell in an Excel spreadsheet, or a line in a log file. Anything that occurred in a point in time."

[Why does Yahoo love Apache Storm? Read Yahoo Talks Apache Storm: Real-Time Appeal.]

The company offers four subscription pricing plans that scale up based on the number of data points and the length of time jKool keeps your data. "We charge for what we measure, which is the performance of time-series events," Rich said. "Right now we're a service, but we may do something on-premises in the future."

Certainly, jKool is far from the only player in this space, but its emphasis on cloud-based analysis of time-series data might help it carve out a unique niche. It sees its target user as a business person, rather than a data scientist or engineer, who wants to find outliers, anomalies, and bottlenecks in streaming data.

"We're focused on time-series data," said Rich. "That's data that has a value at 9:00, and the same entity has another value at 9:05, 9:10" and so on.

An energy company that Rich spoke with at All Things Open offered one potential use. "They're sending engineers with Android tablets into different plants to monitor temperature," he said. "Those plants are hot, and many times the tablets fail."

The company wants time-series data to predict tablet failure, as well as the effectiveness of its plants and engineers, he said.

"They want to know the plant's location, who the engineer is, how long they've been there, the elapsed [travel] time between plants, and temperature of the Android tablet," said Rich.

But how will jKool compete with larger, more established vendors in this space? Rich sees jKool's ready-to-use analytics service as a competitive advantage.

"If you look at Google's and Amazon's offerings, they're mostly frameworks for doing analytics, but they're not the actual analytics itself," said Rich. "Many of the services out there are offering more like toolkits to do this, whereas we've built a full-functioning service."

Ease of use is essential as well, particularly for business users who lack the tech savvy of a data scientist.

"We spent a lot of time building this very friendly query language, so you can go to the dashboard and start talking to your data in very sophisticated ways with very simple language," said Rich.

What will you use for your big data platform? A high-scale relational database? NoSQL database? Hadoop? Event-processing technology? One size doesn't fit all. Here's how to decide. Get the new Pick Your Platform For Big Data issue of InformationWeek Tech Digest today. (Free registration required.)