Enterprise Architecture (I think) Starting from Scratch
March 30, 2016 1:00 AM   Subscribe

If you were starting a brand new small college computer infrastructure, where would you start?

I've been tasked with building a small college infrastructure (Student information system, learning management system, email server, website, security and I hate to add...etc.
My background is instructional design and technology, I've worked in colleges and universities off and on for 16 years. But I've never been asked to build the whole thing from the ground up. I know that ideally I'd start with a needs assessment and requirements, unfortunately all I have is that there will be 500 students and around 60 courses.
Where is a good place to start thinking and conceptualizing the nuts and bolts of this? I'm pretty good at scoping out kind of a "God's eye view" of what's needed. But when it comes to say the first piece of hardware or software to consider, I'm at a loss.

Can you think of any good resources that would help me make decisions? I know this probably sounds kinda crazy.
posted by PHINC to Technology (11 answers total) 8 users marked this as a favorite
 
Don’t underestimate the amount of staff time it takes just to keep a secure mail/calendar/websit/network infrastucture system up and running. That’s before you even think about developing any kind of customised tooling for your institution.

You can go a long, long way with Google Apps for Education & Google’s security is probably the best out there.
posted by pharm at 3:26 AM on March 30, 2016 [4 favorites]


NB You’re going to find a *lot* of vendors pushing 'solutions' for educational institutions: Is there a conference or something you can get to where you can actually talk to the people running these systems at other institutions? That might be where I’d start.
posted by pharm at 3:28 AM on March 30, 2016


Thinking about this as an "enterprise architecture" might be sending you down the wrong path. If there was specific value for you in building something from scratch, maybe because you have unique requirements, or you were looking to get a competitive advantage out of the specific way that you implement your technology - then an enterprise architecture could be a good way to conceptualise what you need.

But it seems that your task should probably be a lot simpler than that. There will be a whole range of packaged applications (Google Apps no doubt being one of them) that other colleges use for their own very similar needs. Many of those packaged apps will be hosted by their respective vendors & made available to you as SaaS, so your own on-site infrastructure requirements to use them will be much simpler - probably just a browser.

So if it was me - I'd be making a short-list of vendors whose products are used by other similar-sized colleges - with a focus on SaaS offerings. Aim to get the basic features established first, and work outwards from there.

Think iteratively - what's the minimum you can run with from day one? What features will you add on in the next iteration, and the next? Go for the easy targets first, and leave the hard stuff & nice-to-have extras until last.
posted by rd45 at 4:10 AM on March 30, 2016 [2 favorites]


Start with Google or Microsoft hosted services for email, etc. They have packages for the education market. I know Blackboard has a hosted option so the other vendors in the course management space must also.

I'd think SaaS is the way to go if you don't have the expertise in-house to maintain secure systems.
posted by LoveHam at 4:30 AM on March 30, 2016


I like what rd5 says here. You shouldn't be building anything from scratch, rather, your job should be researching and selecting vendors and then working with them to get what you need.

Just be aware that someone on your side is going to want there to be a way to log in once on one page and access all of those services. Tell them that will come in Phase 3.

You asked for a good resource to figure this out. Until you have more details, I think you're looking at it. Anyone that you ask about it isn't going to be able to help you until you have more information.

For example, is this college online, commuter, and/or residential? Will you be handling tuition (and other payments: meal plans, extracurricular activity fees, lab fees, etc)? Will the system need to be able to send out blast emails to various groups (prospects, students, faculty, fundraising)? Will the staff need to be able to create and send paper communications? How will student information be collected? Wil it need to handle payroll for faculty and staff?

I think your first order of business is to find out who can answer those questions for you and spend some face-to-face time with them before you do much else.
posted by dawkins_7 at 5:15 AM on March 30, 2016 [1 favorite]


No one builds much from scratch any more: either COTS in your own data center or hosted stuff is the norm -- with good reason! These days I would think very seriously about how much of it you can build out of hosted or SaaS offerings. (And I say this as a guy who is very skeptical of "Put it all in the Cloud!" Pollyannas.)

Last month, the guy who runs Columbia's security group gave an interesting talk about moving .edu applications into the cloud. His premise is that most of the cost is in personnel, so moving to the cloud is maybe 50% likely to be cheaper. BUT you are doing a "greenfield" implementation, so you don't have the baggage of migration.

However, he also talked about how your agreement with the vendors is your sole recourse should there be a problem. Like, can you get your data back if you want to move? What format will it be in? Who accepts/rejects upgrades & patches? Who has back-end access to your stuff? Are there regulatory controls on where your data can be held? How and how well does a given system integrate with your other systems (like getting student rosters & class lists into BlackBoard, or exporting student data to a hosted portfolio offering, or linking up your Admissions CRM with your SIS, etc. etc.)? If they shut down, can they sell off their customer list, or entire operation?

It was a really good presentation, so MeMail me if you wan a link to the summary, or to get a link to where they should be posting up his slide deck Real Soon Now.
posted by wenestvedt at 6:23 AM on March 30, 2016


There will be a whole range of packaged applications (Google Apps no doubt being one of them) that other colleges use for their own very similar needs.

Yes, but be very critical and sceptical; there is a lot of overpriced and low quality software aimed at higher ed, too.
posted by thelonius at 6:55 AM on March 30, 2016 [2 favorites]


Reiterate what thelonius said: edu software can be a real shit-show. Our institute of higher learnin' paid big bucks for a flagship system that does-not-work, and from what I've heard it's pretty common elsewhere in the education field, where money is quietly burnt up and accountability is often shoddy.

Point being, when you shop around be extra vigilant in research.
posted by ovvl at 7:31 AM on March 30, 2016 [1 favorite]


Google Apps will cover a LOT of your needs, including the LMS (Google Classroom) if you wish and even the web site, and it's free for education. It also comes with free ediscovery tools to meet compliance needs. You also get SSO (single sign-on).

Gartner is a "go-to" consultancy in higher ed., and EDUCAUSE is the standard IT organization to belong to.

You can throw up a web site on Google Sites and map it to your domain or go with some kind of Wordpress option, either a hosted service or a VM on a cloud service.

If at all possible, I'd try and keep everyone in the same ecosystem (e.g. Google). Educational institutions can sometimes take a "use what you want" approach and you end up with support and training nightmares.

I think the toughest part will be the SIS. I believe Kuali might be an option there.

If Google Classroom doesn't fit your needs, Canvas is the current flavor of the month.
posted by idb at 10:16 AM on March 30, 2016


Background: I'm an IT consultant focusing on enterprise infrastructure, with a fairly large degree of experience in public education and global megacorp environments.

The following statement makes me extremely uncomfortable:

>I know that ideally I'd start with a needs assessment and requirements, unfortunately all I have is that there will be 500 students and around 60 courses.

This gives me the impression that you are being set up to fail (not necessarily due to malice, but probably due to a lack of understanding of the scale and complexity of what is being asked of you).

A project such as you are proposing requires massive amounts of additional information to make any kind of reasonable headway. Just to speak in broad hypotheticals, you're going to need a good understanding of any physical facilities that will be utilized, the type and quality of services that will be provided to students, and any additional operational requirements that necessarily result in dependencies on IT systems. Tangentially, you also need to be constructing a very thorough understanding of the risk surface that your solutions will be introducing to the institution, and accounting for how (from both a policy and practical perspective) the institution will account for those risks. All of the above will be underpinned by an absolute need to understand exactly how and when funding will be made available and allocated.

You desperately need to work with an IT consultancy that has experience scoping and managing greenfield infrastructure systems deployments. To be especially clear: you need to be working with a firm that can play the strategy game, but also has a considerable amount of implementation experience under their belts. If you do not have the resources or support to involve such experts, then it would likely be in both the institution's and your best interest to back out of this project. Though there may not be any legal ramifications for you in the event of a botched deployment, the ethical consequences are particularly steep in the education sector.
posted by BrandonW at 3:41 PM on March 30, 2016 [1 favorite]


Thank you so much everyone. Again, my faith in Metafilter is made real.
posted by PHINC at 10:35 PM on March 30, 2016


« Older Possible causes of my outbreak of aphthous ulcers   |   Wop-wop-wop laser sound effect? Newer »
This thread is closed to new comments.