According to Forrester Research, “in 2018, cloud computing will accelerate enterprise transformation
everywhere as it becomes a must-have business technology.” Indeed, many science-driven companies across multiple industries are turning to the cloud as a scalable, secure and transformative research-informatics environment for increasing operational agility, lowering total cost of ownership (TCO) and enhancing collaboration.
We see this trend continuing through 2018 as organizations increasingly look to the cloud for a better way to manage the tidal wave of new data types, sources and partnerships engulfing collaborative research today.
We also see the following trends gaining increased traction in the year ahead:
For many organizations in 2018 and beyond, cloud adoption will be a staged process in which scientists accomplish their work using a mix of existing on-premises (server-based) systems along with cloud collaboration using the software-as-a-service (SaaS) model. SaaS manages everything, including data-center infrastructure (e.g., virtualized servers, databases), applications management (e.g., upgrades) and security in a fully controlled, highly available and scalable service.
Most importantly, a cloud-based collaboration system needs to support a ‘hybrid-cloud’ environment in which data flows between on-premises and cloud applications. For example, a design team can start an experiment using on-premises electronic lab notebooks and then transfer information to the cloud for execution at a contracted research organization, which then sends back the results of their work.
This bi-directional integration facilitates a staged movement from existing legacy on-premises infrastructure to an emerging SaaS (or managed-services) environment. An advanced migration strategy will move entire workflows and associated applications to the cloud rather than one server or one application after another.
A 2016 survey of IT and business decision-makers conducted by IDG and published by Informatica found that 74 percent of respondents expect to adopt a hybrid or cloud-only approach to analytics over the next three years. It’s also noteworthy that 52 percent of IT executives report having “cloud-first” policies when making new technology purchases, an adoption stance that is expected to increase to 77 percent by 2019.
When moving to a cloud solution, it’s usually preferable to work with a well-qualified SaaS vendor as opposed to just replacing a local physical server with a virtual machine (VM) in the cloud. Infrastructure-as-a-Service (which requires customers to manage their own applications, security and databases) or a private cloud solution (in which the cloud provider can manage these services at a cost to the customer) could be an intermediate step for customers moving to the cloud for the first time.
However, these approaches may mimic on-premises processes that are not optimized for the cloud, and they are relatively costly. For example, customers will still have to upgrade their operating systems, databases and all applications installed on the VM—and many costs will be charged back to customers because they are applicable only to them (e.g., operating system and database licenses, maintenance and dedicated service upgrades). A cloud strategy built on a proven SaaS/public-cloud foundation can reduce TCO and increase technology availability while keeping client data separate and secure through logical partitioning. In many cases, these are the goals that drive customers to consider a cloud-first approach in the first place.
In 2018, SaaS-based research-informatics networks may also begin to investigate and deploy more ‘cloud-to-cloud’ services that are capable of casting an even wider collaboration net (e.g., the chemical-inventory cloud talking to the enterprise resource planning [EPR] cloud, or the customer relationship management [CRM] cloud talking to the human resources [HR] cloud).
An organization engaged in research informatics might work with an instrument vendor that offers a SaaS product for capturing and storing in the cloud all the data generated on-premises by their instrument. The same research organization might then be interested in transferring this instrument data from the instrument cloud into their notebook vendor’s cloud to support experimental claims. Data transfer such as this will require a ‘cloud-to-cloud’ solution, preferably with single-sign-on capability for an optimal user experience.
With another year beginning, the numbers speak for themselves. More than 43 percent of organizations expect that the majority of their IT capability will be delivered through public cloud services by 2020, and they will access 78 percent of IT resources through some form of cloud by 2018.
Frederic Bost is senior director, cloud R&D, with Dassault Systèmes BIOVIA.