Thomas Barton is an IT engineer at Novartis whose job is to link things together. When Barton wanted to upgrade the company’s middleware-software that connects one piece of software to another--he was encouraged by a colleague to turn to “the cloud.”
In fact, folks in the IT industry are more and more promoting a “cloud first” strategy. Just starting a new business? Don’t buy your own software. Start with the cloud. Upgrading software for your existing business? Step into the cloud. These are folks from Microsoft, Intel, Dell and other IT giants, and their mantra is clear: Waste no more time or money. Go to the cloud first. The message rang loud and clear at the Cloud Slam conference in Santa Clara this week focused on cloud computing and healthcare/life science.
Stats First
Let’s start with some statistics and projections made by keynote speaker, Mark Weiner of Microsoft:
- Data will grow by 4400% over the next 10 years
- By 2015, there will be 2x more smartphones than people
- By the end of 2016, 60% of healthcare organizations will be taking data into the cloud
- Only 16% of healthcare organizations know where archived data resides
- There are 600 million imaging studies done per year
- By 2016, there will be one exabyte worth of medical imaging
With all the data coming down the pike, software engineers say there is no way around it. We must go virtual.
The Growth/Security Conflict
The issue of whether to use the cloud is much like the “open science” conflict I’ve discussed in previous blogs. One must use the cloud to grow in a cost effective and efficient way. Yet one must not use the cloud to maintain the highest security of sensitive data.
“There’s no way our data is going in the cloud,” is a common line industry veterans have heard over the past few years. Yet, business by business, industry by industry, conference by conference IT engineers are pushing back.
The security issue is the main concern and understandably took up the bulk of the time for panel discussions throughout the Cloud Slam conference that were populated with leaders from Microsoft, Dell Boomi, Intel, and other software giants.
So how do these experts address the concern over security and privacy? Healthcare is a heavily regulated industry.
“The cloud can provide better security,” said Microsoft’s Hector Rodriguez. “In fact, we already live in the cloud.”
Microsoft’s answer to the Amazon cloud is Windows Azure, touted as “a cloud for modern business.” Hector went on to explain that without a comprehensive data strategy, employees of healthcare organizations resort to “work around” methods, such as email and social media. Any data shared in this fashion is not secure. Using one system hosted in a cloud platform where data integrity is maintained according to the needs of the organization can thereby improve security.
The cloud is also better for disaster recovery. During natural disasters, such as the recent Sandy storm in the northeast, millions of dollars are spent recovering data centers that were not originally budgeted for. The cloud is not limited by local conditions. When organizations use the cloud, data back up can be done within hours.
So are concerns over security overblown? Panelists at the conference urged healthcare organizations to “push your vendors.” Special cloud platforms can be HIPPA compliant, can be made to work with demands of regulators such the FDA makes on clinical trial data. In fact, it was pointed out, an FDA cloud workgroup has just begun meeting.
An important trend to help business take advantage of the cloud is to consider a “hybrid cloud”, using a combination of platforms, perhaps one internal and one virtual, or two separate virtual systems.
The Novartis Story
Thomas Barton of Novartis began with one project. Their middleware was just taking too long and costing too much. Over the past seven months Barton and his team have changed from enterprise software hosted on sixteen servers internally to eight servers in the cloud hosted by Dell Boomi. Though Barton didn’t share the exact cost savings from this transfer, he did say that the company was now saving 30% for this application.
The Novartis case is an example of a hybrid system. They are converting one project at a time and still able to have all their various software connected.
Barton says skeptics in the company asked, “oh no, where does my data go?” He replies saying that it doesn’t go anywhere. That it is configured and coded in the cloud, but the data is still local.
Novartis deals with sensitive, regulated data, and according to Barton, “everything is validated and qualified.” Disaster recovery has been vastly improved. And, Barton says he’ll see payback within the first year.
There were more success stories. Jason Stowe is the founder and CEO of Cycle Computing and devotes his time to life science customers. He’s convinced that cloud computing has the power to advance science in a dramatic way.
“Researchers can ask bigger questions with new compute technologies,” Stowe said in a breakout session.
In one project, Stowe was able to turn “39 years of science into 11 hours” with his cloud platform. Going to the cloud versus using traditional in house servers, reduced the cost of the project from $44 million to $4,372.
With a genomics project, Stowe’s company did what would normally be 115 years of compute hours in one week for $19,555.
I’m not sure how Jason calculated his numbers, but they’re getting him into some big doors. Cycle Computing serves most of big pharma and some large genomic projects.
Making the Jump
So why don’t more life science companies turn to the cloud? After hearing stories like those of Novartis and Cycle Computing, I expected to see more companies from our industry at the conference.
David Houlding, a speaker and panelist from Intel Healthcare (Intel has a healthcare division?), speculated on the resistance to the cloud.
“It’s been a disservice to the industry that the cloud has been framed as new and risky,” he said. “Actually, what we’re seeing is just the evolution of the data center.” Houlding insists that the cloud is “not an all or nothing journey.”
Other panelists discussed resistance within the ranks of an organization’s IT department. When the idea of moving to the cloud is mentioned and considered, it could lead to jobs lost. Panel members pointed out that overall IT departments are understaffed. And that engineers who have specialized in on-site data centers can go on to become privacy teams, focused on theft protection and other security issues.
“Even though an organization may outsource to the cloud, it is still the one accountable to its customers and regulators,” said Houlding. “IT folks will become auditors.”
The conference was very commercially focused, being sponsored by the speakers from the big software giants. As expected, there was not much push back or skepticism.
The comments of Intel’s Houlding about the disservice of calling the cloud something new and risky make sense on the one hand. The life science industry can be especially slow to adopt new ways that are perceived risky. In fact, I see this as our industry’s core conflict. We have, and must continue to cross new borders to grow and improve. Yet we must be careful not to add more risk.
Sometimes a paradigm shift is necessary. A big jump in conception. Sometimes it is helpful to say that we need to think in new ways. Jason Stowe gives a provocative challenge to investigators, and all of us in the industry. Can we come up with bigger questions, can we think bigger with better compute power?
The question of whether to use the cloud, it seems, is not if, but when.
