The quest for integration nirvana

LAURA WILLIAMS: Our school board has 140,000 students, 8,000 teachers, and essentially every September we are completely reorganizing. We have a quarter-million classes set up. So, one of my challenges is definitely integrating the different applications. Data has to be accurate. We need to make sure students are in the right classes, teachers are in the right schools, that kids are on the right bus. The systems that support don’t have to be real-time, but they have to be timely so the integration is a bit of a challenge. And we have to do it in a cost-effective way, we are public sector and we are always looking for affordable ways to do that. Right now, we interface about 14 systems in multiple different ways so our integration needs are quite complex.

TERRENCE VERITY: In education, as in so many businesses, there are an immense number of challenges, and one of our interesting challenges is (to identify) who the customer is. It would seem that the obvious answer is the student. But when you start to look at the services you offer, it actually extends out beyond the student. The students that might come to college, we need to capture them, (as well as) the people that graduate from the college. We really have to be conscious — and conscientious — that the systems we build (address) not only the immediate student, but all these extended families. Customer focus is really important. The really difficult part is this independence that is there. People always want to build their own applications and do their own thing, and always faster and better than any central unit. We have this push-pull back between reliability and security and the need to support the business in this rapidly changing environment we are in. So not letting silos happen is really important, at the same time as encouraging independence and innovation.

TIM FRY: Our environment is radically different, we have two areas that have conflicting demands. We provide software and services to about 80 customers in Ontario in the retail energy business so they can transfer business data between them, and we also have an internal IT organization. The challenges are quite different. On the internal side, we have to support software developers and they are very demanding, but there are constraints on how much you can control their activities. They are very technical and like to circumvent any type of centralized control. The other side is to ensure that our customers are getting software that works properly and allows them to do their business successfully. The challenges there are dealing with a lot of legacy systems, customer information systems.

CC: Assuming you have all these systems up and running, what is it that drives this need for integration?

TV: And these systems work very well. But in the end, what drives it is service to your customers. For example, at Seneca, we counted them and we had 88 different Web sites, each of them managed by somebody else. A student at the college says, “So, what is Seneca about?” There’s this site, there’s this site and there’s another site. Basically, you have this confused customer. That’s a bad thing. You realize you have to change your structure. You have to put that together so you can do a better job of customer service.

LW: In our sector, there is a growing interest in accountability and reporting. We have a huge responsibility to report to the Ministry of Education on our student performance. What’s also driving it as well is that we built these systems for operations, and they are silos, and now we are trying to get essentially what is executive reporting around what is happening in the organization. So all of sudden we have to integrate across multiple systems in almost real-time.

TV: Post-secondary is exactly the same thing plus accountability for good decision-making. As our president said, and I give him full marks for saying this, “Lot of data, but not much knowledge.” So how do we do this?

CC: Knowledge management is a term you see everywhere, but what does it mean to you?

TF: To us, it’s primarily technology knowledge and problems we have faced in the past. There are always different layers of data that eventually become knowledge.

TV: It’s not that we don’t have the data, we have to know what questions to ask, we have to build the integration structures. Most important is delivery. Reports don’t do anything unless they’re in place where they can be acted on in terms of accountability and decision-making.

LW: We have a similar challenge in that we have a window of opportunity in terms of the students that pass through our doors. We can’t wait until they are well through our system to find out if they are struggling. We need really timely data not only about where they are, but whether the interventions are working. It’s really important to know within a month whether or not a certain program is succeeding.

TV: For both our organizations, there is a question of accountability in terms of doing the right things for students. We have just put in a new vice-president of access and success — access to make sure students not just get in, but stay in. For that, you need the right kind of knowledge and the right kind of relationship with a student that says, “We are going to work with you, we are not going to leave you.” So, IT gets to build a lot of great systems that do that. A Web-based appointment book is huge because students need access to counsellors, so we built a really great piece of technology for the student and the people who need to help that student.

CC: Do you have a problem with applications proliferation?

TV: The Web has encouraged users not only to develop them, but implement them and get them to customers. But it’s dysfunctional, and pulling that all back together, where the data is protected and secure, is a challenge.

LW: I think also there is a tremendous resistance to architectures and the perception is that architectures slow you down. And it’s true, because architectures have corporate value, but (they) also have local value (to business units). So, can a business unit get up and running really fast? There needs to be this type of corporate ownership that says we are going to do architecture, we are not going to build silos, and it is going to slow you down, but it is a corporate (objective).

TF: My experience is that it is a very noble goal to have this top-down architectural standard that you enforce . . . and it’s fairly easy to define them.

TV: But you have no choice, you have to do it. You have to clearly state the capabilities, say it’s supported by the executive of the organization, and that this is the way the organization runs, and it’s in everybody’s best interest to run this way.

LW: I think you need that corporate buy-in into architecture. I’m not sure it’s about saving money. It’s about risk and security, doing integration after the fact is enormously expensive and difficult, and sometimes not possible.

CC: The industry is pushing service oriented architecture. Is that helpful?

TV: I’m not so sure it’s an industry push. What we see is this environment and the Web, which really is a set of open protocols. There’s standards for the server and for data transport and for the browser, and yet, these are really good things that say these are standards you can write to. SOA and Web services really are standards written by agencies that say this is how we see things working. For the most part SOA and Web services are standards we can understand and buy from them and implement ourselves. The question is, is it going to make a difference to us? Yes.

LW: We’ve always had this problem of integration. And SOA is a helpful paradigm and a new way of thinking about this problem. You are not going to solve the whole integration problem, but it will give us a different lens to solve part of it. What’s really interesting is I know two organizations that have essentially implemented pieces and one was really successful and one wasn’t — with the same technology. It’s not about the technology, it’s much more about the structures you have to put into place.

TV: All of us have a big investment in infrastructure and programmer knowledge and the responsibility to run these production systems on a day-to-day basis and not take them down and say, “We’ll be back in a year with a whole new architecture.” You have to keep going. But it does give us the opportunity to try and fix things, to develop new things. We know how to do this, so let’s get the training and let’s see how this feels. For example, when high school students apply to college, they don’t apply to a college, they apply to the Ontario College Application Service. One of the things that organization wanted to (implement) was a transcript exchange because students need transcripts sent from high schools, and transcripts are a really important part of this, and there are EDI standards for these transcripts. But the data comes in many different forms from many different institutions. What we need to do from a low-cost perspective is find an architecture that looked at this process differently. So we suggested looking at it as a Web service so the data moves with its description. So we built that and we now have the experience to build an architecture we can use for the rest of our environment.

CC: Where are you with Web services?

TF: We have some Web Service implementations that fall in to some degree an SOA paradigm, and in allowing people in applications to talk to each other in a very efficient way, but for us, Web services and SOA is very much customer-driven exercise.

LW: Internally, there is more control so you might be using more Web services there. The example I give is the credit card processing. Everybody is willing go buy that service inside the organization. I don’t know if we are ready to go outside our firewall and buy that service from another organization.