Four Steps to a Big Data Strategy
Big Data is playing a huge role in medical research—some even believe it will be instrumental in finding a cure for cancer. Though in its early stages, harnessing the power of Big Data obviously has the potential to change medical research in a major way.
The National Institutes of Health apparently agrees. This week, the NIH announced funding for the establishment of six to eight investigator-initiated Big Data to Knowledge Centers of Excellence. The funding will be for up to $24 million per year for four years.
“The centers will improve the ability of the research community to use increasingly large and complex data sets through the development and distribution of innovative approaches, methods, software, and tools for data sharing, integration, analysis and management,” Scientific Computing reports.
The centers will also provide another much-needed resource: training for students and researchers to use and develop data science methods.
The NIH will also look outside the traditional biomedical areas for expertise: The NIH will form interdisciplinary teams that include data scientists. If you’re interested, there’s a free webinar for prospective applicants on Sept. 12 from 3 p.m. To 5 p.m., EDT.
This is part of the NIH’s Big Data to Knowledge initiative—BD2K for short—announced in December.
SOA: Is Failure an Option?
Joe McKendrick at ZDNet recently shared two pieces about high-profile companies using a service-oriented architecture: Twitter and Netflix.
The Netflix piece gives us a peek at how important resiliency and failover become for a mature SOA. For Netflix, a subscription-based service, “failure is not an option,” McKendrick points out.
On the other hand, at Twitter, “failure is always an option”—in fact, the company allows for failure, according to Jeremy Cloud, leader of Twitter’s Tweet service team. Cloud is overseeing a less mature SOA, and he offers three “lessons learned” for companies undertaking SOA.
One of the lessons that surprised me was how hard integration and integration testing is. Wasn’t SOA supposed to simplify integration? That’s not what happened for Twitter.
“Planning your integration test and strategy is very hard,” Cloud said. “We’ve learned, unfortunately the hard way, that integration on SOA is actually really hard… When you have dozens of services that all have to work together, it presents a pretty complex picture.”
How hard? Testing is manual and tedious and is still an “unresolved problem,” according to Cloud.
Enterprises Embracing APIs
A recent survey of 140 enterprise IT professionals found that more than 43 percent of organizations already have an API program in place, with an additional 27 percent planning to launch one by next year.
What’s driving the adoption of APIs? Primarily it’s mobility, with 71.9 percent citing this as a reason for starting an API program. Other top drivers:
- Partner connectivity (69.1 percent)
- Cloud integration (67.6 percent)
- Enabling internal developers (66.9 percent)
- Fostering external developer ecosystems (55.4 percent)
The survey was conducted by Layer 7 Technologies, which offers API management and mobilization tools.
