Yesterday -- oh, o.k. the date... 14th January 2009.... we held our first Cloud Roundtable. The topic - “Current and Future Trends in Cloud Computing” with a focus on Biotech / Life Sciences / Pharma.
Our guest speaker was non other then Dr. Werner Vogels. He is the CTO and VP of World Wide Architecture at Amazon. Dr. Vogels and his team are responsible for Amazon’s cloud architecture that we and others are using on a daily basis. Among his many accomplishments, Infoweek selected him as CO of the year last month.
We had a seating limit of about 35. Forty two people signed up and it appears that 48 walked through the door. For a while there I thought I would have to pull the airline trick and offer 2 free tickets to the Cloud09 event to anyone that gave up their seat. Luckily our co-sponsor Pfizer, was kind enough to supply extra chairs. I thought Werner would draw a crowd but I didn’t imagine we would get a 120% turnout. Not bad for our first event. I put this down to a pent up demand for information about Cloud Computing in San Diego, in particular within the biotech / pharma community. I’m sure having Dr. Vogels on the docket helped a bit.
Werner started off by asking the crowd what they would like the talk to center on. He could cover the history and why Amazon decided to build the cloud architecture (1 vote) or give examples of how others within the life sciences are using cloud computing (everyone else’s vote). He started off by talking about a few biology and chemistry applications that some are using and how they scale. Things like BLAST among others. He also pointed out that there is a growing number of public data sets hosted on AWS: ENSEMBL’s Annotated Human Genome Data, NCBI’s GenBank, PubChem and others.
I got the sense that many were pleasantly surprised that Amazon is hosting these data sets. I think there were a number of people in the audience that suddenly had some ‘Ah ha!’ moment when they realized the power of having data sets located close to their compute and storage infrastructure. I expect the AWS Public Data Sets will continue to grow both in the number and type of data available. Watch this area folks, for many killer Cloud Application are all about data.
There were a few questions around Amazon’s security model. Werner explained how security for cloud computing is an extension of the retail security model that supports the Amazon Store. In fact, one interesting comment was that he considers the Amazon Store just another customer of AWS. Thus, if you have your compute infrastructure on AWS, you are using the same security model as the world’s largest retailer. Pretty cool, when you think about the amount of data Amazon must host. There is a caveat though, you have to take responsibility that you are locking down your server through access control lists, proper setup of ports and port mapping, patching your AMI, etc. The same sort of thing you would have to do within your own data center. Kurt Messersmith, Amazon’s Cloud Dude <my title, his card is in the other room> put it best. Security is really about trust. Amazon has a robust security model, the question is really do you trust Amazon to hold your proprietary and confidential information?
Surprisingly, there was only one question on validation. This is a hot topic in the Pharma space. The FDA has expectations in regards to computer validation. A company in theory can be shut down if they do not meet those “expectations”. In a very simplistic sense and one that any auditor reading this is bound to yell at me “But, you are missing <fill in the blank>!!”, - system / application validation simply demonstrates that your company is following good IT practices; security and access controls, a robust change control methodology, staff and user training, system logging and documentation describing how you are doing all of this. There aren’t any technical reasons an application residing in the cloud could not be validated. There is just a lack of experience and thus comfort level on how to achieve validation.
This is an area we at Cirrhus9 are are working on. When I spoke with Werner afterwards I brought up the validation issue. He said that Amazon is very interested in helping pharma meet industry requirements. This doesn’t mean that Amazon will carry out the validation, but they are willing to have a discussion on ways to make validation possible. This whole are is worthy of another blog post and a discussion on what is needed to achieve validation. Cloud computing is here to stay and we will make it work for Pharma.
After about an hour and a half, the questions briefly hit one of those quiet moments. Werner took this time to show some non - biotech uses of AWS. One of my favorites, was Animoto. From this site, you can create mini-videos mixed your own images and sound track. Or use music available on Animoto’ site. What is interesting about this company is they set a service level agreement with their customers that generation of the video will take at most 5 minutes. They pull this off by scaling up and down the number of servers carrying out the video encoding. Nice. For biotech, that means if you have a compute intensive calculation you could set up your environment so that you can always have the calculation completed in a fixed amount of time. Project managers are just going to love that. Animoto is a very nice use case for the scalability of AWS. Go ahead and try Animoto. A 30 second video is free for the making.
To wrap up - our 2 hour roundtable ran over by about 20 minutes. I was remiss about keeping Werner to a time schedule, but the audience was so interested I don’t think too many minded. Werner demonstrated that the power of Amazon’s architecture (scalability, network, compute power, services) provides a beautiful pallet to build applications upon. Applications that would be very difficult to build within your own company data center. And the best part is Amazon is always looking for new ways to add additional functionality at the infrastructure level.
As always,
Quello è tutto per oggi!
-P
