- Hybrid cloud storage environments are a reality for healthcare organizations seeking to digitize their health IT infrastructure. However, many entities are challenged when deciding which workloads and applications are best suited for the cloud, and which are better off remaining on-premises.
ClearSky Data CEO Ellen Rubin told HITInfrastructure.com that hybrid cloud storage is here to stay for the foreseeable future. While some organizations are eager to move parts of their IT storage infrastructure to the cloud, many are still hung up on which data sets belong on which type of storage solution.
“Healthcare organizations are living in this hybrid reality for the foreseeable future so they need to decide how they want to handle the portfolio of infrastructure,” Rubin stated. “That’s what entities are trying to figure out right now.”
“Some workloads will never go to the cloud,” she continued. “Some things can go and may keep primary production data in their data center but they’ll use the cloud for archival or backup. Organizations will be comfortable about certain workloads being built in the cloud and living there permanently.”
Rubin typically sees all of these instances present in every healthcare cloud storage client. Organizations are faced with overcoming general distrust of cloud storage, while making informed decisions on which data lives where.
“The healthcare industry is unique because it has such strict compliance and security concerns,” explained Rubin. “Healthcare companies are all trying to adopt the cloud.”
“They have a huge infrastructure that has always been in their data centers and under their control. They have to manage and run the data center while supporting very heavy workloads for EHRs and analytics and everything else involved with running the business.”
Data protection is the number one concern healthcare organizations face when determining the best way to store their data. Rubin finds that health IT infrastructure presents a very large, expensive, and complex set of data challenges at the storage level.
“Healthcare organizations want to move to a cloud environment that saves money and comes with exciting and compelling features,” said Rubin. “They are also trying to adopt digital transformation strategies, which means they need to develop and release software more quickly.”
“Entities are looking to the cloud model because it gives them faster delivery and more agility,” she continued. “Organizations are sometimes stuck in a box as they try to innovate and add to their storage infrastructure.”
IT departments must appease patients and clinicians with faster and more innovative IT infrastructure. However, these departments are also hesitant to let go of their current infrastructure because data stored on-premises is under their control.
Rubin said she encounters this dilemma with every healthcare organization considering a cloud solution.
“Healthcare is divided on their opinion of the cloud for data storage,” said Rubin. “Some are eager and excited to move to the cloud while others are struggling with how to make cloud storage work in a way that doesn’t put the data at a higher risk.”
In addition to cloud security concerns, organizations are also faced with justifying rebuilding workloads and applications that are running smoothly in their traditional data center environment. Rebuilding workloads and applications for the cloud is a strenuous undertaking that makes a migration strategy particularly important.
“What you see happening is security teams weighing in on what workloads can safely move to the cloud,” Rubin explained. “It’s a negotiation trying to figure out which workloads an organization would consider moving to a cloud environment. Are there any new projects the IT department would be willing to start with in the cloud? Are there any workloads we can migrate off that feel less risky?”
“It’s hard,” she continued. “A lot of the time cloud service providers are being asked to help the customer make their transition to the cloud and work with them while they figure out their approach.”
Organizations need to decide where to store their data based on the nature of the application and the nature of the data, Rubin maintained.
“Data that is patient identifiable or PHI is sensitive,” explained Rubin. “Even organizations that are heavily using the cloud may hold off on patient data and chose to move internal IT-types of workloads where data is sensitive but not extremely compliance oriented.”
Entities typically start by moving low-end or smaller workloads and applications to build up confidence before migrating more sensitive or robust workloads. Those workloads that contain too much data may not be worth the trouble to migrate.
Entities often start with newer datasets that grow once they are deployed in the cloud.
Generally, data migration is determined by where the application wants to run. The standard rule is that larger applications that need to scale up on the compute side are a better fit for the cloud.
“Things the customer feels don’t need to live in their data center anymore, like VMware environments, are moved to the cloud because they take up a tremendous amount of resources,” said Rubin. “The VMware environment is already virtualized and organizations are willing to consider that it doesn't have to sit on physical gear.”
“It’s easier to move a VMWare environment to the cloud than workloads that haven’t been virtualized,” she added. “There are characteristics of the application that make it more cloud friendly or less challenging to move.”
Once it’s decided what workloads and applications are moving to the cloud, the actual migration needs to take place. Rubin finds that healthcare organizations tend to underestimate how difficult it is to migrate to the cloud.
“Initially, organizations underestimated how hard it would be to deploy real IT infrastructure in the cloud,” said Rubin. “They didn’t think about data migration, latency, or rebuilding applications for the cloud. These are the things that organizations have been learning the hard way over the past several years.”
The cloud is much different than what IT administrators are used to deploying in their legacy infrastructure environments.
Organizations should start with a strong understanding of the skills and processes involved before and after the migration occurs, including management and maintenance requirements.
“Organizations are now much smarter because they have tried different cloud tactics and solutions to varying degrees of success,” Rubin explained. “They realize that managing the cloud is managing infrastructure, too, even though the data is not in an organization’s data center and they don’t have to scale or handle data on-premise.”
The network requirements of data migration also present an obstacle for many healthcare organizations.
“The ability to just move data in terms of the bandwidth that would be available for a large, heavy movement to the cloud or from the cloud is where things get ugly,” Rubin said.
“An organization has to be willing to invest to upgrade and scale their network. In some ways, it’s a short-term problem. If you have a dataset and you’ve moved it, then you’re done because everything that happens with the data from that point is in the cloud.”
Rubin advised that entities think about the migration processes holistically by involving the security team and the finance team from the beginning of the cloud deployment and migration process.
“The cloud providing capacity on demand is a change for the finance department. The security team also needs to be involved in the process to ensure that organizations don’t sign deals with vendors that don’t adhere to their security protocols,” Rubin explained.
Trials and proof of concept to get comfortable with the technology and seeing how it runs is also necessary for a successful cloud deployment. Organizations need to be sure that users are able to easily retrieve data from the cloud and that workflow is the same or better than before the cloud deployment.
“In healthcare organizations especially, making a change brings a lot of questions,” Rubin concluded. “Showing that the cloud solution works before it’s fully deployed gives users confidence in the technology and reassures that their workflow will not be disrupted by the change.”