Moving to the Cloud Means Offloading Risk
One of the drivers behind cloud adoption is risk aversion (see related). Many are under the mistaken assumption that premises-based solutions are less risky than cloud alternatives, or that risk is a constant. Premises-based solutions certainly offer value in control and customization, but the risk averse are gravitating toward cloud services.
Like the hypothetical frog that fails to detect the increases in water temperature before it boils, the dangers of increased risk also often go undetected. Over the past 15 or so years, premises-based solutions have become a risky proposition. Technology improvements introduce new capabilities and efficiencies, but also transfer responsibilities and risk from manufacturers to customers.
Premises-based solutions, particularly when compared to the cloud, are becoming untenable in terms of risk. Let's recap what's occurred over the past few generations of enterprise voice to make this so.
Starting in the '90s, the PBX was a pretty low risk proposition. All of the hardware engineering was done by the manufacturer. These systems were difficult to over-provision because the appliances ran out of slots and ports before running out of processing capacity. The endpoints were tightly integrated and connected over dedicated networks. Parts and licenses were sourced through authorized channels to ensure proper configuration and installation.
In the 2000s, things began to get more complex and the risk burden began shifting. VoIP meant the communications traffic now traversed shared networks. This introduced the need for voice prioritization techniques such as VLANs and tagging. To power endpoints, customers bought new switches with PoE--often before the standard settled. SIP promised interoperability, but it was initially rough--customers had to sort out which providers and endpoints were compatible. The session border controller (SBC) came soon after Internet connectivity, to mitigate new kinds of vulnerabilities. The iPhone not only pushed mobility in general, but forced enterprises to expand their voice-capable networks to include Wi-Fi.
Around the same time, the industry started moving away from appliance architectures to standard servers. The change was hugely beneficial to the industry in terms of price and performance, but shifted more responsibility to the customers. For example, component upgrades such as additional memory were no longer a simple call to the dealer for a SKU. Instead, customers had to sort compatibility matters. Even worse, the certified engineer that did show up to fix things no longer had the scope or skills to manage customer-provided hardware.
Engineering also shifted to the customer with software-based solutions. End users now bear the responsibility for ensuring adequate hardware capacity. The manufacturers are happy to sell as many licenses as the customer wants--extensions, contact center agents, conferencing resources, etc. However, it is up to the customer to ensure the configuration meets the documented requirements.
Then came virtualization and even bigger challenges. Now the IT staff are responsible for managing the separate requirements of the hypervisor maker as well as the UC application vendor. Onsite certified technicians can't even reboot the system (instance) without customer involvement.
That's just voice. Now factor in all the complexities of UC--unified messaging and directory synchronization, video and presence gateways, mobile clients, and APIs. The risk proposition has consistently been shifting toward the customer over the past 15 years. Although the modern UC solution is supposed to work like the PBX did before, the complexity and points-of-failure exponentially grew.
The complexity of the current environment becomes abundantly clear when a problem needs to be isolated. IT staff are expected to troubleshoot clients and devices, wired/wireless networks, the virtualization and server environment, the application configuration, SIP provider services, campus networking and configuration, headsets, Bluetooth, licensing options, recording solutions, scheduling integrations, and whatever else it might take to solve the problem.
Oh, and all this risk has increased while staffing levels (and training budgets) decreased.
Certainly the cloud doesn't eliminate risk, but does shift considerable portions to the provider. The beauty of a service is that the underlying details become someone else's problem. CIOs no longer need to be concerned with server maintenance, staffing coverage, and virtualization. A good SLA: Priceless.
The business of on-premises communications has become a risky proposition.
Dave Michels is a Contributing Editor and Analyst at TalkingPointz