Are We Ready for Adopting Cloud Computing?

Cloud adoption in the pharmaceutical industry is steadily growing despite the posed challenges. Why? Because by embracing the cloud, pharmaceutical providers are able to access scalable storage solutions designed to handle massive amounts of digital information without overstretching IT budgets. Companies like GlaxoSmithKline have reduced their cost by 30% by adopting cloud based solutions and other players like Pfizer, Eli Lilly, Johnson & Johnson, and Genentech have already begun experimenting with cloud computing. This article looks at the recent successes and provides insight into enabling pharmaceutical workloads on cloud platforms.

The leading pharmaceutical companies have started hosting complex genetic sequences, clinical trials, and biomarker data in the cloud. Cloud based computing is ideal for data storage and analytics of the vast amount of data scientists and biologists generate in research. Eli Lilly is using the cloud to provide high-performance computing and storage to hundreds of its scientists, enabling it to reduce the fixed IT costs without compromising with IT services.

Pharma and Cloud: The Challenges

• Data Set Size: The typical pharmaceutical workloads generate or consume huge amounts of data. The data store is generally collocated with the computation resources. Moving data from an enterprise location to the cloud and back is prohibitively expensive in terms of bandwidth and time. With a 4 Gbps network speed, it would take more than 2 hours to move a 4 TB data set. In data centers and enterprises, the data throughput is typically around 100 Mbps. The cost and actual time for data transmission could undermine this whole approach.
• Regulatory Compliance: In the pharmaceutical industry, there are strict regulations relating to personally identifiable information, where private data can be stored, how and what applications can access that data. One of the challenges is ensuring and proving regulations have been met. Cloud services come with a basic set of capabilities and building blocks to meet these needs. Any solution that has to go through strict due diligence, auditing, and certification processes is a high-cost activity.
• Security Concerns: Another aspect of regulatory compliance that needs to be addressed is the various security levels required of pharmaceutical data: Physical Security, End-point Security, Storage Security, Transport Security, Deletion Security, etc.

Leak of company confidential information must also be addressed. In a multi-tenant environment, its critical to ensure the application accessing data is trusted (signed and certified) and isnt running in a shared environment accessible by another company.

Cloud Applicability and Approaches

To commence, its important to conduct an assessment to figure out which applications and data can be movedkeeping in mind the criticality of the information that needs to be moved.

Data Migration

An easy way to start is to move smaller data footprint workloads. Once the cloud stack is tested and acceptable, larger workloads can be moved to the cloud. The initial input data, currently within the firewall, has to move to the cloud. Once the data is secure on the cloud, it will remain there for future workloads and analysis as the cost for bringing it back is high and counterproductive.

Storage

Any workload and associated data should be checked for data access patterns. If its raw data from a study, the data access could be within a second and may recur frequently in a day. If its patient images, access time may be a few seconds or minutes, but the frequency of access could be once in a few weeks. If data stored is associated with a research query, then the access time can extend to hours and frequency may be 1-2 times per year. With this in mind, tiered storage can be created that has costs corresponding to the retrieval rate and responsiveness required.

Security

Its possible to address security requirements of the pharmaceutical industry via a few approaches:

• Private Cloud: Utilizing a private cloud platform can meet regulatory and confidentiality needs. Consumers of the service can utilize the self-service to launch workloads and use the reporting framework to determine the expenses related to the execution. Maintainers of the service can utilize the multi-tenancy model to reduce licensing fees, standardize on hardware, and achieve better economies of scale. Pfizer has been using cloud for its R&D projects and has adopted a virtual private cloud that builds an invisible wall around the companys servers within a shared or public cloud. These servers, hosted by Amazon, allow the company to extend its firewall and other security measures to the cloud. The use of thousands of offsite servers has helped Pfizer compress its computation time from weeks to hours.
• Public Cloud: The least sensitive and lowest strategic value workloads can be run on public clouds in an effective manner. If researchers around the world need access to data, the public cloud can act as a data center which can be accessed by researchers everywhere. Eli Lilly, a global pharmaceutical company, moved its newly commissioned research workloads to a public cloud rather than investing in buying extra servers for their data center.
• Hybrid Cloud: It is possible to retain sensitive data within a private cloud and use a public cloud to satisfy excess computing requirements that dont require access to sensitive data by using a virtual private cloud. This establishes a secure channel between the private and public clouds. Provided that the public instance does not persist data, this model can be used to create a hybrid cloud. Even if some sensitive data is required to be shared with the public cloud instances, a secure channel can be used to ensure transport security. The pharmaceutical companies conducting large-scale clinical trials can adopt hybrid clouds and continue maintaining storage for operational data inside their data centers. The companies can continue to run intensive clinical analysis computational processes on the cloud. This hybrid approach delivers scalability, while reducing computing costs at that scale on a pay-per-use model that a cloud computing environment offers without exposing mission-critical data to external vulnerabilities and threats.

Major Sources of RoI

Where the cloud can help pharmaceuticals reduce costs include:

• Business Alignment: With a cloud platform, it is easy to see the exact cost of a workload. This breakdown leads to better alignment of the IT resources with the business goals. If a workload costs a lot to run but doesnt yield a corresponding return, it can be reviewed for effectiveness. The R&D team can focus on reducing the costs of such instances. The management team can make informed decisions on product development strategy based on this information.
• Data Access: Once the data is on a cloud, it can be reused by multiple applications. This design can create data reuse not possible before. If the data is interesting to partners, allied research organizations, or even competitors, then it can be licensed out to these parties on a flat fee or utilization basis.
• Creation of Services: A cloud technology platform certified by regulatory bodies and proven to scale to match pharma workloads is a valuable asset. This technology can be licensed or provided on a hosted basis for the entire industry. A conglomeration of pharmaceutical companies can set up a joint working group to realize this vision.

Conclusion

With the industry needing to improve productivity, maximize existing investments in IT, monetize data assets, build a quicker road to product development, and focus on personalized drug creation, the cloud is no longer just an efficient way to manage IT functions, but the beginning of new business model for the industry. 2012 is the year pharmaceuticals will embrace and reap the benefits of cloud computing. Are you ready?