By Sramana Mitra and guest author Shaloo Shalini
SM: Interesting! So, you basically leap-frogged the previous generations of on-premise technology and went straight to cloud computing because that is what makes sense right now, that is where the good stuff is in terms of time to market and rapid adoption, right?
DM: Yes. Earlier, I mentioned that we are headquartered in Lafayette, Louisiana. We are 40 miles off the Gulf of Mexico. Five years ago, during my sixth month on the job in Schumacher Group, the state of Louisiana was hit by Hurricane Katrina. In my seventh month on the job, we were hit by Hurricane Rita. There were those two storms in Lafayette. Luckily our headquarters and our community were spared by these storms. But I saw the difficult situation my peers in southeast and southwest Louisiana were dealing with in terms of the susceptibility of their data centers. I recognized that our data center too was vulnerable to storms. At that time, we didn’t have a good business continuity and disaster recovery plan in place. So, that was a major driver for me to sign a contract for CRM solutions with Salesforce.com.
Once we got ourselves enabled with Salesforce.com, we found such huge value that we kept adding more features and functionality to it. We have been very much in line with Salesforce.com. We have taken their product out as a platform and what I do now is ask, ‘Hey, can we do this inside of Salesforce.com? If it’s not a right fit, let’s see if there is another solution available out there that we can use. Or is there an on-premise, off-premise solution?’ We do our typical product selection and evaluation to meet our business needs. At the end of the day, nine times out of ten, we bring up cloud-based solutions. Today, what we are working on, on a go-forward basis, is where we really feel those services such as Amazon S3, EC2, and all other similar services out there, including those that are competing against Amazon, have reached a stable point. Now we are looking at some of our in-house, on-premise data center solutions and gearing toward moving some of those solutions over to the Amazon services.
SM: You mentioned that you have ramped up quickly and bridged the gap in the application structure of the organization over the past five years by using public cloud-based applications. What about infrastructure? Are you using Infrastructure as a Service (IaaS) today; is that part of your strategy? What is going to be the evolution of IaaS?
DM: I think that the overall prices decrease with the Infrastructure as a Service solution. We use more of Platform as a Service (PaaS) solutions these days as opposed to Infrastructure as a Service (IaaS). We have been ahead of the curve and taken care of things such as integration and single sign-on. We are integrating the services we run inside of our data center with these cloud-based solutions. Many of these services have moved more into IaaS in the past two years. We are re-evaluating our contracts for the appliance devices we have in place. It’s more than likely that we will move those services into cloud. The reason for this move is again to reduce dependency on our data center footprint, which is important for business continuity.
SM: I see. One comment I hear quite often from CIOs is, ‘I don’t want to be in the data center business and moving to the cloud.’ What is your reaction to that?
DM: You are talking to a guy who had a UPS failure the day before Christmas, and the day after Christmas the utilities went out. I had to bring my data center down and had to hard restart twice in a five-day window over the holidays. None of my IT guys really got to enjoy the holidays. So I would agree with that comment.
SM: I think almost everybody we spoke to shares this sentiment right now.
DM: Yes, it was an eye-opener for the network operations team, no? I know we like having an Exchange Server mail setup inside of our own environment. At the same time, we also use Google e-mail. I have 2,500 Google e-mail licenses for all of our providers. I know we like having an Exchange server on site, but why does it have to be on site? We are all scratching our heads and having a lot of conversations with the network operations teams about what needs to be inside the site. As we continue to grow as a company, we are thinking about whether we can deliver [. . . ] services that have better redundancy, more focus on security, and more such features with, say, cloud-based and shared models scenarios.