Higher Level Decisions — One University's Cloud Strategy
A Q&A with Tom Hoover
Aging hardware, failing servers, interrupted e-mail service — these were some of the painful elements that initially prompted UTC's serious consideration of a cloud migration for office and productivity solutions. But the ultimate result was a major shift to a much wider cloud strategy that offers wins in many areas for better services and economy. Associate Vice Chancellor and CIO Tom Hoover explains why most of his institution's traditional IT services could eventually be hosted in the cloud.
Mary Grush: Was there a particular moment or turning point when your institution began looking seriously at moving certain IT functionality to the cloud?
Tom Hoover: Yes. It all started for us about a year and a half ago, when we were beginning to deal with some ailing, failing servers and hardware in our Exchange environment, where we were at the same time running out of disk space. We decided on an initiative to migrate to Office 365 — in the cloud.
For us, it just made sense. As an institution, you examine your core motivation for change. Buying hardware that you must replace every five years no longer made sense. Our strategy needed to change.
One striking factor related to the issue of planning for hardware replacements is that we tend to get "one-time money" at the university. You now find that at a lot of institutions: It's been difficult to get recurring funding.
So we began to move e-mail to the cloud. Our current faculty and staff are now in the cloud. We are in the process of moving our personal and departmental shared drives, in addition to SharePoint, to the cloud.
All of this has provided us greater reliability. We don't have the hardware and out-of-space failures that we had in the past. It's also given us better support, and it's available on all of our devices.
It's a true win-win for us. For services now in the cloud, we don't have to replace associated, aging hardware that's coming to end-of-life. So we're addressing that problem while at the same time improving in other areas, such as economy and service.
Grush: Of course, you didn't stop with e-mail.
Hoover: That's right. For example, we're now moving our listserver to the cloud, and we're trying to take advantage of more hosted solutions where it makes sense. In fact, most of our services are at least beginning to move to the cloud, and many are already there.
For us, it just makes sense. It's easy to manage cloud-based services, you don't need to worry about the hardware, and you basically pay as you go.
Grush: Does the pay-as-you-go factor make your migration easier?
Hoover: Yes, you don't necessarily have to make a huge initial investment. You may just invest incrementally when you need more space with the hosted service.
Grush: What about looking at the horizon? Do you have an idea of your next steps, very far into the future?
Hoover: Yes. In the future, there will be very few applications still on premise. These will be certain specialized applications that cannot be hosted.
Looking now at how hosted solutions are moving in general, and the leveraging and economies of scale you can foresee for your own institution, it just doesn't make sense to plan to leave most traditional applications on campus for very long into the future.
Grush: What is one of the more important areas you are considering now, for migration to the cloud?
Hoover: We are now doing a proof of concept for putting our disaster recovery and business continuity in the cloud. Think again about that more general picture on the horizon: At other universities ERP systems like Ellucian, Banner, and PeopleSoft are now in the cloud. With major business systems running in the cloud, there's a big piece already in place for your DR strategy.
In our case, we are almost done — extremely close to finished when this Q&A comes out — with our business continuity study. We are defining which core services will have to stay up — our ERP, our Web site, and associated authentication servers. Those are the kinds of things that we've considered most carefully in our proof of concept in the cloud.
It's notable that the process of a proof of concept includes, among other things, going down a list of services to look at all the components of your strategy. Checking through that list, we found that a very large proportion of these services are already in the cloud. Just as examples, we have TouchNet, which is how we handle our payment processes through Banner — in the cloud. We use OrgSync here at the university, in the cloud. And of course, I already mentioned Office 365, which is in the cloud. And as I said, we're moving our listserver to the cloud. We are moving the way that we manage Macs to the cloud. There are so many examples. You end up thinking, "There isn't that much that isn't in the cloud already!"
Grush: Thinking about your users, are your hosted cloud services going to be highly usable and customizable enough that you'll be able to serve users well?
Hoover: I'll repeat here that I think we are giving them greater services in the form of greater reliability.
Case in point: A couple weeks ago, one of our core servers on campus had to be re-booted on a Sunday. This caused a campus Internet outage for a while. But, most users were not affected, because their Exchange e-mail is connected out through the cloud. The users didn't notice a thing. This would not have been the case in our old on-premise Exchange environment.
So in the case of many future cloud services, the users won't need to know or care that these services are not on campus any more. But they will be getting better, more reliable service.
And in terms of customization, we will be able to spin things up as needed. There will be the potential of offering VMs, whether that's in AWS or Azure, that we can spin up literally in minutes. Think about that. Traditionally, with everything on campus, we would have to look at whether we had enough VMware space and power before we could bring up a VM for someone. In the cloud, it's going to be easier and quicker to spin up these custom environments for people. It's more flexible for us, which translates to greater customization for our users as needed.
Grush: Does this logic apply to research projects?
Hoover: Sure. Considering server installations for any research project, you finally begin to think, "Why spend a half million dollars on servers — hardware that you'll have to replace in five years anyway — when you can pull them up as needed with Azure or AWS and have as much compute power as you need?"
Grush: What's the nature of security and exit strategies in most cloud environments?
Hoover: In some respects, your environment will become more secure when you are passing certain responsibilities on to the cloud provider. For example, with Microsoft Azure, you have FERPA compliance already built in. That's welcome assistance for the institution.
In terms of an exit strategy, you still own your data and could, if really necessary, move that data to another cloud provider or even back to campus. In that respect, there is always an exit if you actually needed it.
Hopefully, though, you have already looked carefully at the services you are trying to provide and have analyzed what makes sense before your move to the cloud. How you will be allocating your resources either to the cloud or to your local data center environment should be well thought out, of course.
Grush: What do you observe about the effects of the cloud, from the point of view of the IT department itself and how it functions? Can it move IT's focus toward higher level decisions?
Hoover: It can allow you to focus IT staff on what your core competencies are as an institution — whether you are a master's institution or an R1, or anything else, you can potentially redirect some of your personnel resources to address broader goals.
Each campus has to look at what works best for them. For some institutions it has been very difficult to hold on to data center people because of pay scale relative to the industry in general. In that case, the institution may be able to redirect some of these relatively higher-cost personnel and leverage their skills in new ways for higher-level tasks.
Ideally, taking advantage of the cloud will allow us to shift the focus of the IT department in general towards creating a better student experience that will serve the university's goals for retention and student success. That's the ideal — actually getting there is the more challenging part.