Symyx

One of the more interesting talks at 2010 Symyx Symposium in Barcelona provided an overview of cloud computing. Because the speaker’s organization doesn’t allow its employees to talk in print, I opted to summarize some of the points from the talk. This first entry defines some terms around cloud computing and offers advice on how to evaluate the different types of cloud offerings available. I’d love to hear about whether and how you would consider hosting data in the cloud.

The true business problem most life science companies are facing today is the externalization of the scientific workflow. To achieve the ultimate goal of designing, making, and testing compounds, organizations need a way to work across complicated layers of interaction between globally dispersed departments, acquired companies, and external research partners.

In our personal lives, many of us are already using and trusting cloud computing, as this video whimsically illustrates.

We are networking on Facebook, storing our photos on Flickr, and even banking online. Could these tools become enterprise workhorses? Maybe, but achieving this will require organizations to take a hard look how they work—and what they expect from the software tools they invest in.

Determining which type of cloud computing you want depends on what level of risk you are willing to take with your data. 

• Public clouds: Data, processing power could be at any time on any IP address on any machine in any country. Public clouds are therefore problematic for any data that might have associated commercial liability or that must be closely audited.
• Private clouds: Cloud-based technology is used to run a virtualization that is secured behind a private firewall. Private clouds are ideal when you need to know where your data is, as they provide the audit trails required when dealing with intellectual property (IP) or personally identifiable information (PII).
• Hosted: Traditional servers or virtualization behind a private firewall. In a hosted model, vendors simply provide and maintain services off-site for customers.
• Internal: Traditional software ownership model where a company purchases software and maintains it behind its own firewall.

 The next question is what type of service you need. Today, you can choose from several types of cloud-based “as a service” offerings. 

• IaaS (infrastructure as a service): The hunt for cheaper CPU cycles, storage, and infrastructure costs
• UCaaS (unified communication as a service): The hunt for cheaper communications infrastructure costs, such as VoIP
• SaaS (software as a service): The hunt for cheaper infrastructure costs and software utilization
• BPaaS (business process as a service): The hunt for cheaper total cost of ownership of a business process, which requires a realistic understanding of risk and technology.

Obviously, working in the cloud requires you to have a cloud strategy. You can have a floaty, cirrus-based cloud strategy or a deep, cumulus-level strategy that’s really integrating it into key work. Our Symposium speaker argued that the best approach is a cumulonimbus strategy that covers all the bases. And this means asking the right questions of your organization and potential providers. I’ll present these in the next entry.