Cloud Considerations for Medical Device Makers

Thinking of connecting your medical device to the cloud? Here are a few things you should know.

November 11, 2015

5 Min Read
Cloud Considerations for Medical Device Makers

Thinking of connecting your medical device to the cloud? Here are a few things you should know.

Dana Good

It's an exciting time in the medical device industry, with new “cloud-connected” systems emerging every day that offer significant new value to patients and medical service providers. These new device systems—and even some implanted systems—do more than collect and display sensor and patient data; increasingly, they reach out wirelessly to the cloud for data storage, computation, and directed alerts.

These new connected systems are having a real impact. When network-connected medical devices provide patients and providers with instant information, a world is created where important information such as an alert from a patient's insulin pump, heart monitor, or fall detector can be quickly relayed to a physician or member of the family. The computational capabilities in the cloud allow device manufacturers to offer big data services to their clients even if the client has no computing hardware. And a well-designed cloud system can operate with an availability level of 99.999% or more (a key metric for reliability measurement).  

But with capability comes complexity. Cloud service providers are making a real effort to reduce this complexity, but there remain significant design challenges and a steep learning curve.  This article addresses some of the areas to consider when developing a medical device system using the cloud connectivity model.

What is the Cloud?

We hear and see the buzzword “cloud” everywhere, but most people haven’t yet worked with the cloud directly, and there can be some confusion about what it means.

Fundamentally, the cloud is a set of software services and resources that can be remotely accessed over the Internet. The power of the cloud becomes evident when taking advantage of “metaservices” such as automatic scaling, load balancing and scheduling to create a system that runs with high availability, locality, and performance at scale.

Cloud service providers often offer connectivity to their services and resources in the form of a software development kit (SDK) that allows accelerated systems development and deployment. Additionally, cloud providers offer dashboards, browser plug-ins, and other tools to simplify and allow automation of system management. This is a cloud platform.

What Does the Cloud Offer?

The on-demand, scalable resources the cloud offers include the following:

  • Secure storage: Reserved storage on remote servers.

  • Compute power: Virtual machines, including web servers, with load balancing and load-based scaling.

  • Database as a service: Distributed, low-latency, very high availability, automatically scaling, SQL, or NoSQL, hosted by the cloud service provider.

  • Authentication services and security token providers.

  • Specialized DNS routing optimized for locality.

  • Distributed, scaled, internode communication facilities.

These services are integrated with common development and deployment platforms (.NET, Node.js, Windows, Linux, Mac OS, iOS, Android). They allow deployment and data synchronization across multiple geographic and logical zones to provide locality and reliability. They allow infrastructure monitoring and automation through dashboards, programmatic interfaces, and e-mail/SMS alerts.

Moving to the Cloud

Before making the leap to the cloud, here are a few things medical device companies should consider.

Leverage Existing Technology. Large enterprise systems have been steadily moving to the cloud for years, preparing the ground as they go.  Any company is looking to develop or rerelease a product in the cloud would do well to consider using the well-developed SDKs now being offered by the larger cloud service providers such as Amazon and Microsoft.

Define Nonfunctional Requirements. Nail down your nonfunctional requirements (e.g., reliability, availability, security, privacy, performance, scalability, and cost) before developing the system architecture.

Use Best Practices. Cloud-based systems need a cloud-based architecture—the most critical part of a successful cloud project. Information on best practices in cloud architectures is now widely available and should be followed to ensure success.

For instance, Netflix built its systems to fall back to functional backup servers during cloud outages; this is why they stayed up and running so successfully in 2011 when Amazon Web Services had widespread outages.

Build in Security and Privacy. The need for security and protection of privacy can’t be overstated.  Even so, backfilling secure authentication strategies or secure database access after beginning development is a common and risky approach. Avoid this mistake by considering and designing for your security needs up front.

Partner with a Cloud Services Provider. Choosing a provider of cloud-based services may be the single most important decision to your new system. Fortunately, the landscape has thinned out significantly in the last few years. The field is now dominated by Amazon Web Services and Microsoft Azure, with Google bringing up the rear. A number of specialized cloud providers also exist, such as Rackspace, IBM, Virtustream, and Verizon Terremark.

Many of the providers’ offerings are very similar. If you are considering just Amazon Web Services and Azure, one consideration would be the desired underlying software development platform. Azure is somewhat (though not entirely) Windows-centric and might be a more familiar environment to Windows developers, while Amazon Web Services is the “longest and strongest” player in many ways, has more offerings, and has been running many types of servers for quite a long time.

Final Considerations

Monitoring and Managing the Cloud. Managing a cloud installation will require some new skills and tools.  Happily, once the work of coming up to speed is done, IT staff will be less in the business of keeping individual machines running and more in the realm of managing systems for reliability and flexibility.  The monitoring, management, and automation tools offered by cloud service providers take time to learn but offer enormous power.

Cost. Cloud services are an operating expense rather than a capital expense; cloud service providers, after all, are responsible for upgrading and replacing the hardware and software resources. The largest players, Amazon Web Services and Microsoft Azure, are roughly similar in cost, and the more specialized providers can cost less. Don’t forget to give these smaller cloud providers a look if your needs line up with their offerings.

Don't miss the BIOMEDevice San Jose conference and expo, December 2-3, 2015 at the San Jose Convention Center.

Dana Good is a software engineer at Stratos

[image courtesy of JUST2SHUTTER/FREEDIGITALPHOTOS.NET]

Sign up for the QMED & MD+DI Daily newsletter.

You May Also Like