Athanasoulis: The business of Harvard Medical School is research. ... Similar to many industries, there is a culture that requires that, for IT to be successful, it has to be meeting the needs of the users.
We have a particularly interesting situation. I call Harvard Medical School the land of a thousand CIOs, because, in essence, we cannot mandate that anyone use central IT services, cloud services, or other things. So that sets a higher standard for us, because people have to want to use it. It has to be cost-effective and it has to meet their business, research objectives.
We set out about five years ago to start thinking about how to provide infrastructure. Over time, we've evolved into creating a cloud that's a private cloud at the medical school.
We've been able to put in place a cloud that, number one, has user participation. This means that the faculty have and the researchers have skin in the game.
They can use the resources that are made available and subsidized by the school, but if they need additional resources, additional computing power, they're able to buy it. They actually purchase nodes that go into the cloud and they own those nodes, but when those notes are idle, other people's work can run on it. So they buy into the cloud.
These folks are not very trusting of central IT organizations. Many of them want to do their own thing. In order to get them to be convinced that they ought to participate, we told them, "You buy equipment and, if it doesn't work out for you, you can take that equipment and put it under the bench in your lab and set it up how you want." That made them more comfortable. But, not a single time has anyone ever actually come back and said they were going to take back the equipment.
In essence, it's building the trust of the researchers or the business clients, if you're in more of a business environment, getting them engaged in their requirements, and making sure it will meet their needs.
... Personal relationship is a part of what it's about. We had to make sure that we weren't seen as just a black box that they had absolutely no control over. That was step number one.
Then we also had to make sure that it was very much of an iterative process. We would start with one folk's needs and then realize there were certain other needs.
... We started out with a relatively small cloud initially. Once people saw the value, they began to adopt it more, and it's really starting to have a snowball effect, where we are growing by orders of magnitude.
... People are moving from the giant project, two- to three-year implementation cycles to, "Let's take a chunk, see how it works, and then iterate and moderate along the way."
Skilton: What's illustrated [at Harvard Medical School] is this need to move to more continuous-release or continuous-improvement type of life cycle. This is a transformation for IT, which may be typically more project-cycle based. It's a subtle difference, but it's one that is fundamentally changing the way you would offer an incrementalized service as opposed to more of a clunky, project-based, traditional waterfall approach.
We're seeing software as a service (SaaS), due to the economic conditions, taken quite seriously now, particularly targeted at specific business processes, but also starting to become potentially more mainstream. Clearly, with Salesforce.com and others like that, we are seeing that starting to accelerate.
... We're starting to see utility computing becoming much more common mainstream, so that it's no longer a fad or an alternative to mainstream. We're seeing that sort of consistency.
Demonstrate success early
Athanasoulis: It's always easier to show someone something that's already working and say, "Do you want to hop onto this bus" than to say, "We're going to build this great new giant infrastructure, and just trust us, it's going to work great. So, hop on board now, before anyone has even seen it or tried it out." It's having the ability to let people walk before they run. Come on and try it out. If it doesn't work for you, so be it, but you also have demonstrated successes that people can point to.
... The CIO at Harvard Medical School, John Halamka, had the vision to start this. It started with his initial vision and going to bat to move from everyone from doing their own thing and setting up their own infrastructure, to creating a cloud that will actually work for people.
He had the foresight to say, "Let's try this out." He went to his leadership, the dean and others and said, "Yes, we're taking a chance. We're going to spend some money. We're not going to spend a huge amount of money until we prove the model, but we're going to have to put some money in and see how this works." It was a very interesting communication game.
Peyret: From an enterprise architect (EA) point of view, we should ... determine what are the elements that can migrate to the cloud, different types of cloud. Then, we should try to evangelize. The EA should be in between business and IT. That's a good place to make a right choice and mitigate risks and choices.
... The EA should participate to establish and negotiate what I call the business service catalog, something that will be an extension of the ITIL service catalog, which is very IT-based and IT-defined.
Something that is missing currently within ITIL V3 is how to deal with the business to define the service and define also the contract in terms of cost and of service level agreement (SLA). But, it's not only the SLA. It's broader than that. That's something that's missing at the moment. Most of the EAs are not participating in that.
... The business service catalog is the next step. We have heard in enterprise architecture about business capabilities. We talked about that business capabilities to help develop business architecture.
A missing link
We have also heard SOA. There is a missing link in between -- the business service catalog. It's a way we will contractualize. I like very much the fact that you said, we are contractualizing, but with flexibility. We should manage that flexibility. We should predict what that flexibility means in terms of impact. Perhaps that service is not valuable for other parts of the company.
That's where I think that EA and the next step for EA will take place. SOA is not an end, and the next step will be the business service catalog, which we will develop to link to the business capabilities.
Isom: The catalog of services would be great. I think we need to be careful about that catalog of services, so that it doesn't become too standardized.
As I mentioned earlier today in one of my presentations, you want to be careful with that standardization, because you do want to give people some flexibility, but you need to manage that flexibility. So, you need to be careful. We need to be careful with the catalog of services that we offer, but I definitely think that it is a new way of thinking, when it comes to the role and capacity of IT.
It's a new way of thinking, because along with that comes service management. You can't just think about offering the services. Can you really back up what you offer? So, it does introduce more thinking along those lines.
... The enterprise architect would be the one who would provide that enterprise view and make sure that anything that we do is thought out from a holistic perspective, even though we may actually start practicing on a smaller scale or for a smaller domain.
A good practice would be to involve the enterprise architect, even though we may start with a specific domain for implementing the cloud, because you've got to keep your eye on the strategic vision of the company.
... As far as what's driving cloud as a solutions strategy is the need to improve business performance. If we can get solutions that will help drive business performance and business sustainability, the cloud is a good place for that.
... You can't produce cloud solutions in a vacuum. You won't get any consumers. So, it's a great venue for cloud providers to work with business stakeholders to explain and explore opportunities for valuable services.
Athanasoulis: Defining the service with the users is the first clear step, and obviously getting the requirements from the users, particularly in an organization like our medical school, where they have choices and they don't have to use the systems.
... As IT leaders, we all know that there is now a marketplace. The public cloud is available to folks. People can get on Amazon EC2. They can get on to these various clouds and they can start to use them. That forces us to have compelling cloud offerings that are more cost effective than what they can go get out in the public sector.
... We view the public cloud as an extension of the private cloud to the degree that there is consistency of virtual machine definitions and to the degree that we can make a node on the public cloud look exactly like a node on the private cloud and make the same databases available there.
If someone has the money, they want the capabilities, say 10,000 processor hours or 100,000 processor hours, whatever it might be, between now and this deadline three weeks from now, and they are willing to spend the money, wouldn't it be great if transparent to them, they just spend up to $100,000, $200,000, whatever their budget is, and let this stuff go from our private cloud out to the public cloud. What a great solution that would be for folks.
... So, having this balance of bringing in an IT specialist, the enterprise architect, to define the requirements in joint-step -- back to the dance with the customers -- was really what allowed us to be successful.
A new question
Skilton: The portfolio needs to be put in place, but it also needs another set of service management investment tools to control data distribution, compliance, or access and security control, and things like that.
I detect a worry about whether I can outsource that. Do I need to do something in-house? What do I need to spend money on? Because that's a block, and people need to understand that.
... What we are seeing with clients now is that they are over the initial infrastructure as a service (IaaS), platform as a service (PaaS), SaaS, and business process as a service-sort of conversation. They're now asking, "What cloud services do you do?"
What they mean by that is that they need to see your cloud security reference model. They need to see your cloud services model. They need to understand the type of services that you can offer into a portfolio and then the types of service catalogs that you can interact with them.
They then make a decision. Does that need to be on-premise, can it be out in the cloud, or is there something as a hybrid? They're on that page now, and there is a strategic planning process starting to evolve around that.
Athanasoulis: You want to iterate and you have to have a vision of where you are going.
If you're taking a car trip and you're going to drive from here to Ohio tomorrow, we know where we're going, we have our map, we start to drive, but we might along the way find, that the highway is clogged with traffic. So, we're going to go around over here, or we are going to take a detour.
Perhaps, somewhere along the way you say, "You know what, now that we have been learning more, Ohio isn't really where we wanted to go. We actually want to keep on going. We're heading right out to Colorado, wherever it may be." But, you have to have a vision of where you are going.
Then, to keep things from spinning out of control along the way, it's really important to know the potential factors that might lead to things starting to fall apart or fray at the edges. How do you monitor that you have the right capacity in place? You don't want to sell something to everyone and then find six months into it that you're way oversubscribed and everyone is bitter and unhappy, because there isn't the capability that they expected.
Isom: The IT department should be more focused now on providing information technology as a service. It's not just a cloud figure of speech. They are truly looking at providing their capabilities as a service and looking at it from an end-to-end perspective.
That includes that service catalog and includes some of the things you were talking about, how to make it easier for consumers to actually consume the services, and also making sure that the services that they do provide will perform, knowing that the business consumers will go somewhere else if we don't. The services are just that available now. You really have to think about that. That shouldn't be the driving force for us, providing IT as a service, but it should be a consideration.
Peyret: What I wanted to recommend is that you should evangelize your IT person to act as an IT service. What does that mean? That means that you should recommend to them to contractualize their service, to express and establish, through the business service catalog, including some pricing aspects. Within the enterprise, where you have some funding and no problem about funding, you should contractualize. That's absolutely key to make the adoption of cloud, any type of cloud, easier. That would be more or less transparent.
Enterprise Strategy Group's Lab Validation Report on TSM for Virtual Environments. See why TSM is one of the preeminent backup solutions for VMware and other virtual servers. more
IBM Tivoli Storage Productivity Center can help reduce storage costs by enabling integrated management of storage assets, performance and operations from a single, web-based console. It also integrates with IBM Cognos Business Intelligence for reporting and analytics. more
This EMA paper gives insights on why storage matters for cloud and what's the advantages of storage virtualization for cloud. It reviews IBM’s software defined storage infrastructure solution and highlights the competitive differentiator for IBM's SmartCloud offering. more
The next generation of simplified backup administration dramatically improves scalability and efficiency. Experience how IBM’s advanced interface for Tivoli Storage Manager enables consolidation, intuitive problem resolution and integrated team collaboration. more
IBM Tivoli Storage Manager (TSM) provides a turnkey solution to a range of data protection issues. This complimentary ESG Lab Validation focuses on key improvements in the TSM platform that drive greater scalability, efficiency, and availability in storage management. more