Cloud Computing | October 2012 Digital Edition

By Banding Together, Some Districts Save Big On Cloud Computing

A major statewide initiative in North Carolina is showing how a consortium model can minimize risks for districts and help them exploit the advantages of cloud computing.

This article, with an exclusive podcast, originally appeared in T.H.E. Journal's October 2012 digital edition.

When Ed Chase joined Edgecombe County Public Schools in Tarboro, NC, this year as technology director, he faced an IT disaster.

The networking electronics, procured between 2003 and 2005, were "past end of life." The wiring in the district was never done properly, and over time its condition has only worsened. The district, or local education agency (LEA), is running Nortel 100 megabit switches, whose software hasn't been updated by the company since before it declared bankruptcy at the start of 2009. No security patches. No upgrades. When somebody runs a device with a network card that the switch doesn't recognize, the switch reboots itself.

"Imagine the frustration of having a network that is not well documented, that is old, and that freaks out when somebody brings something in and plugs it into the network," Chase says. "All of a sudden you have 10 classrooms where everything goes down."

Chase dreams of the day when, laden with E-Rate discounts, he can begin anew, starting at Layer 1 and doing a complete rewiring of the district. But the dream comes with a twist. He wants to move as much as possible of the IT infrastructure and instructional technology to a services model. Key aspects of district technology will arrive directly from the cloud, and he won't have to worry about network management, content filtering, or any of a dozen other IT burdens currently on his small tech team's plate.

"If we can have a service with four-nines (99.99 percent) or five-nines (99.999 percent) availability while being supported by staff (here on site), yeah, that helps tremendously make sure that our clients--teachers and students--have what they need when they need it and have it be reliable," Chase says.

The problem is that Edgecombe isn't large enough to attract the interest of leading infrastructure cloud providers, which puts the district--like most other school districts--at a disadvantage when shopping for services and negotiating with vendors.

But because Edgecombe is lucky enough to be located in North Carolina, Chase is going to get his wish. He intends to exploit a major cloud initiative being refined in the state and involving every one of its districts.

5 Lessons for the Aspiring Consortium
The North Carolina Education Cloud, or NCEdCloud, a multiyear project that started in late 2010, is designed to provide structure, processes, and assistance that districts require to identify IT needs, potential service providers, and even motivation for making the leap from on-premises to cloud-based operations.

NCEdCloud doesn't deliver the services itself; it acts as a hub for all the interactions that need to happen to put cloud services in place. NCEdCloud is funded through a combination of state and local education agency support, but eventually the services are expected to attract enough district participation to become self-funding.

The many lessons it has picked up along its evolution are applicable for other IT-oriented entities--including regional service centers, consortia of districts and small colleges, even large districts--that want to band together to shift key portions of their own operations to the cloud, without doing the heavy lifting of running those services.

Lesson #1: Deeply Understand Districts' Needs
NCEdCloud compiled a 300-plus-question online survey to capture the details for each district of its IT infrastructure, including costs, service models and vendors, and operational and instructional goals. To help the districts fill it out, the NCEdCloud team pre-populated as many answers as they could with data pulled from state sources; they asked specific questions and avoided open-ended ones; and then interviewers spent between two hours and a full day meeting with technical, financial, instructional, and administrative people in each district to fill in gaps.

Every district completed a survey, providing a rich data vein to mine in developing needs, requests for proposals, and other building blocks that make up the plan for moving services to the cloud. "Our work is very much based on what we learned in that process," says Phil Emer, program director for the Friday Institute for Educational Innovation and head of NCEdCloud.

Lesson #2: Focus Goals on Pain Points
Based on the survey finding, Emer says they focused on "the top priorities that cause pain in some way to the (LEAs). We can get our arms around the service. And there is a mature market for it, meaning several service providers are prepared to compete for that business."

The out-of-the-gate goals for the organization are organized around five pain-reduction strategies:

  • To transition district infrastructure to a cloud service model, with services such as content filtering, firewall security, and single sign-on;
  • To deploy a statewide learning infrastructure, such as a repository for maintaining learning objects and digital content;
  • To modernize business operations systems, starting with moving HR and financial operations off district-run IBM i-Series AS/400 hardware;
  • To form a support team that can help districts with their migrations; and
  • To create the systems needed to deliver anywhere-anytime access to the shared learning infrastructure for all teachers and students.

Lesson #3: Clearly Define Your Scope of Service
There are a lot of moving parts in a cloud consortium and being absolutely clear about what your group does and does not do is critical to establishing trust with members.

NCEdCloud doesn't host any of the infrastructure itself. (For a contrast to this model, see the IlliniCloud business model.) It contracts with other entities not only to deliver the actual services but also to act as service managers for each cloud service being provided. For example, in the case of content filtering, the role of service manager is filled by MCNC, a nonprofit organization that also handles delivery of broadband to schools. MCNC in turn is in charge of negotiating the contract and managing the relationship with the vendor that will actually deliver the filtering service, in this case Zscaler.

So just what does NCEdCloud itself undertake? It performed the survey work, of course. But also, it handles planning, engineering, and designing. As Emer observes, "With the connectivity project [see "Bandwidth First," p. xx] we kind of stopped at the edge of the school. We drilled a hole in the school and stuck fiber in. But with cloud-based services you're now going into the school, down into the user level." That brings a new layer of requirements: "Someone has to collect the information, figure out what's possible, figure out what vendors are in place, and figure out how we're going to pay for it," he says. "And so we act as sort of a planning and design shop."

The team also provides ample guidance to the individual districts. Members of the leadership team can send experts into a district to help it handle the migration of a service off internal infrastructure and onto an external service provider.

Likewise, when a working group needs to be pulled together to tackle a specific cloud service evaluation project, the NCEdCloud leadership has deep enough understanding of its constituent districts to pull together a "dream team," with representatives from the state's eight regions.

Lesson #4: Make the Service Opt-in and Motivate First Adopters
Neill Kimrey, director of the Instructional Technology Division at the state's Department of Public Instruction, has a line he likes to use when describing what success looks like for NCEdCloud: "We want to hit 80 percent of the need for 80 percent of the LEAs." The goal is to provide viable cloud solutions to hit most if not all of district needs. "Do we think that we're going to provide an identity management solution that 115 LEAs are going to use? Probably not. But we are going to do our best to provide a top-notch one that will be less expensive than if they do it on their own," he says.

Ultimately, the math potentially adds up to about two-thirds of the districts choosing to adopt any given cloud service through the consortium--and about a third of districts not wanting to use the cloud for a given service. As Kimrey explains, "Maybe they invested a lot of time and staff development. Maybe they just started a long contract, and they have four years (left) on a five-year contract."

So, in order to get momentum going for any given cloud service, the project team puts mechanisms in place to motivate districts. For example, the first 20 districts to eliminate the use of their AS/400 computers--which every district has--and shift their financial and HR work to a cloud service won't have to pay for the first year of service--NCEdCloud will pick up the bill.

Edgecombe's Chase plans to be the first in line for that. It isn't just his 5-year-old hardware that's driving him to get rid of it. "It's having a place to put it that's secure, that has good power, good cooling." He has none of that right now. "I feel like we would be a really nice test case."

Lesson #5: Be Prepared to Educate Cloud Vendors on K-12
Emer has found that districts in his state will frequently choose to work with a small technology vendor located down the road over a proven service provider with a national reputation. In fact, he adds, when certain large vendors are mentioned in K-12 gatherings, "the audience will hiss." Emer has made it his mission to ask those vendors, "Why do these people hate you so much?" While they frequently answer, "Because we're too big…" Emer digs below that to understand the real reasons. Often, he'll figure out they don't treat K-12 customers the way they want to be treated, with dedicated support, marketing and sales, and K-12-friendly purchase agreements. That's frequently a measure of the smallness of most K-12 purchase deals, he notes.

Emer's ulterior motive is to make sure there are no surprises in the bidding process. He wants to know who will respond to a given RFP and whether they're really up to the job. So that means winnowing out the smallest local vendors who couldn't sustain the kind of rollout necessary to support 115 districts; and it's educating the biggest vendors about how to work with K-12 customers and why they should bother. "We don't want to kowtow to vendors. But sometimes we say, 'Okay, great, we're trying to aggregate demand here. Let's assume you get a bigger slice of the pie, what does that mean for us? What are you willing to do?"

When the RFP is finally released and winners chosen, Emer says the ideal is to have two vendors on that list--"one with 60 districts and one with 55, and I want them fighting to get more market share. You can't always do that but that's definitely a good model."