Decision Points on Cloud Communications for Large Enterprise
Three different case studies reveal varying reasons for choosing on-premises or cloud-based deployments, where cost, resiliency, and flexibility are among the key factors.
Cloud communications gets lumped in with the rest of "Cloud Computing," though it is really a different prospect. The obvious difference from of the rest of cloud services is the real-time nature of communications. Payroll systems, customer Management databases and even storage may require bandwidth and fast response times, but nothing is as unforgiving to a slow network connection as real-time voice and video services.
Case Studies: Two Large Enterprises
We often refer to "cloud based" as an externally managed solution and "on premises" as an enterprise IT managed solution, even if the majority of the "on-premises" equipment is at the corporate data center. The actual business question is whether an enterprise should run the PBX with its own resources or outsource it to a service company. Here is a look at two enterprise customers and the decision points that drove their solution to a self-managed communications system (Robert Half) or communications as a managed service (Gatwick Airport, UK).
* Nikki Roberts is the Global Network Services Manager for Robert Half. Many No Jitter readers may remember Roberts from last year's Enterprise Connect Panel Session, "Is There a New Model for Enterprise Communications & Collaboration?" Robert Half runs more than 10,000 handsets on a ShoreTel platform out of their own data center, using internal staff to manage the solution.
* John Keegan is the Director of Communications Solutions Business at Fujitsu UK. Fujitsu was engaged by Gatwick Airport to deploy 1,100 stations initially, and eventually scale up to 7,000 stations using Cisco's Hosted Collaboration Solution. Front-end helpdesk support is provided by another Gatwick Partner, XChanging. Presumably, Gatwick's communications infrastructure is about as hands off as an organization of that size can get.
Gatwick looked at both on-premises and cloud solutions. John Keegan recalled, "Gatwick knew they wanted to take the airport forward and offer more advanced unified communication technologies. They weren't sure whether they wanted it on premise vs. a cloud solution, and they weren't sure exactly what technology." They ended up with an externally-hosted solution in order to offer services not just to their corporate users but to concessions across the airport.
Robert Half selected the ShoreTel Platform in 2005. Nikki Roberts recalled, "A cloud solution wasn't even on our radar in 2005." However, last year Robert Half upgraded to the current ShoreTel version and still chose to deploy in their own data center. Nikki Roberts's perspective is: "There isn't currently a cloud-based solution that covers every single aspect of the phone system in the way we use it. Do I feel that for certain parts of my organization a cloud-based solution would be OK? Yes, but for the core part of my organization that may be another couple of years down the line, and that's just to do with functionality."
Avaya's Differing Perspective on Large Cloud Deployment
Another view comes from Bob Camel, Senior Marketing Manager for Managed Services at Avaya Operations Services (AOS). AOS sets a threshold of 5,000 stations for multi-tenant solutions.
"The private cloud dedicated for a small or a medium-sized business is called COS Express (Communications Outsourcing Solutions Express)," said Camel. "We have some customers that are much larger. When you start getting that big, if you have 5,000 users, it requires a lot more customization, whether they are bringing it in-house or not." The customized offering is called COS (sans "Express"). Still, Avaya's position is that even with a larger solution that needs to be customized for each company, Avaya Operations Service is providing the same "UC as a Service" value.
"We can do a full OPEX model; pay by user or pay by port," Camel said. "Customers like that in today's tough economic world. It helps them adapt to new technology because they are not training or hiring people and it's all OPEX spend."
For very large enterprises, externally managed communications is seldom presented as a lower-cost solution in terms of hard dollars. Even in the 2,000-seat range, the TCO was lower for an on premises solution, according to an Enterprise Connect series of mock RFPs coordinated by Brent Kelly, Marty Parker, and David Stein.
It would follow that even larger enterprises would have a tough time justifying the outsource model's economy of scale--since economies of scale will also be applying to ever-larger in-house deployments. In other words, whether you're an outsourcer or an enterprise deploying in-house, just because you need two support staff to manage 1,000 (seats, subscribers, whatever "units" you are supporting), it does not follow that you need 14 support staff to manage 7,000 units. If nothing else, you at least have better coverage for vacation days!
Nikki Roberts agrees that at such a large scale, the value of managed services is not about cost. "There are definitely some advantages to having someone else house your equipment, but when it comes to a large-scale phone system, I am not sure that cost is one of them," said Roberts. "It depends on where the business wants to see the costs. I haven't specifically researched it, but I feel like the break-even point would be about the same."
Resiliency and Business Continuity
Managed service providers point to the level of network resiliency and the support infrastructure as the real cost differentiator. Bob Camel cited Avaya's Matrix Management Platform (available only to AOS customers), their four U.S. and 11 global data centers dedicated to AOS support, and the staffing that are behind AOS. "The support matrix is near real time. We have over 750 dedicated communications managed services engineers as well as the normal Avaya technicians. Our Global presence is 24 X 7; for a customer to try and build a platform like matrix would be very expensive, tens of millions of dollars."
Fujitsu's John Keegan has a similar view of the network on which Gatwick's Cisco HCS deployment is supported.: "A key part of having this sort of service is that we deliver it highly available within a single data center, then duplicated across two data centers. The level of resiliency that you can build into it, you could not justify doing on premise."
Nikki Roberts had a more situational viewpoint when asked about system resiliency: "There's more than one way to answer, because resiliency is based on failover. No solution is perfect. Disasters don't discriminate: Is the disaster happening to me, or is it happening to my service provider?"
These comments reminded me of a client a few years ago that transitioned to a cloud-based contact center solution. They are a loan servicing company with nine regional locations across the United States, each with a team performing the same customer service tasks for the individual region. If any site were to fail, the other eight could absorb the workload of the other site. With their own replicated call servers, the resiliency they could build into their own network could easily exceed a cloud provider's. And the calls don't need to be redirected anywhere they normally wouldn't go. A large enough enterprise is bound to have potential failover sites that are already part of their infrastructure.
Next page: Scalability