No Jitter is part of the Informa Tech Division of Informa PLC

This site is operated by a business or businesses owned by Informa PLC and all copyright resides with them. Informa PLC's registered office is 5 Howick Place, London SW1P 1WG. Registered in England and Wales. Number 8860726.

The Challenges Associated with Managing Private Clouds

The cloud! The cloud! It's all we seem to hear about nowadays. We see cloud commercials on TV, billboards on the freeway and huge signs in airports. The Unified Communications industry was initially slow to shift to the cloud, likely because of the real-time nature of the applications and the difficulty in delivering these types of services. But UC has caught up in a big way, and there's a plethora of options for organizations looking to deploy UC in the cloud today.

One big decision for organizations regarding cloud-based UC, or UCaaS, is whether to go with public or private cloud. My research has shown that although there are a number of high-quality public cloud offerings, the majority of organizations want to deploy only a private cloud, or more likely in some sort of hybrid environment.

On the surface, the difference between a public cloud deployment and private cloud seems negligible. In one case the infrastructure is on the enterprise premises, in the other case the infrastructure is hosted somewhere off premises. Seems like a straightforward difference. However, there is one other major difference that could have a significant impact on the success of the deployment, and that's the management software.

As I've pointed out in many blogs and Enterprise Connect panel discussions, UC management is broad and very difficult. Some of the areas of UC management, such as monitoring, provisioning and calculating mean opinion scores (MOS) have gotten better over the years. This has helped improve the overall reliability and stability of UC systems. In fact, when deployed correctly, UC solutions today can provide better overall uptime and quality than legacy systems, largely due to the maturation of fault and performance tools for UC systems.

One area that still causes immense amounts of pain is in the area of on-boarding and provisioning, particularly in large enterprises or multi-vendor environments. There are many issues related to provisioning users, including:

* User and Group Diversity: Diverse business groups and individual users each have their own specific requirements, meaning hundreds of user profiles are required.

* Geographic Complexity: The large number of branch locations, remote offices and teleworkers create a management nightmare. While centralized management would be ideal, not all systems or functions can be provisioned centrally "out of the box."

* Dial-Plan Migration and Provisioning: Managing dial plans is one of the tasks where the level of difficulty and complexity is underappreciated. Migrating dial plans from legacy systems to multiple UC systems can often take months to do and is highly user dependent. If done successfully, no one notices. When done poorly, everyone notices. Without a good management tool, this process can literally take months to complete.

* Multi-Vendor Provisioning: Each vendor has its own unique syntax and implementation of protocols, creating challenges running across complex network architectures.

* Disaster Recovery Planning: Creating continuity of communication services requires more than just network redundancy and multiple call servers. In a disaster situation, user profiles, dial plans and other policies need to be migrated over as well.

These aren't the only challenges related to provisioning, but they're certainly some of the more notable ones. So what needs to be done to help overcome these issues? Well, here is where I'll throw the problem back at the vendor community. In the telecom space, most solutions have management functionality integrated into them. For example, Cisco's Hosted Collaboration Solution (HCS) has the VOSS management tools integrated right into it. This means telcos can modify or build new flexible and standardized processes around the set of tools there instead of having to cobble together a number of disparate tools.

However, enterprise solutions aren't quite there yet. Now in fairness to the vendors, "private cloud" solutions aren't nearly as mature as the ones for public cloud and have focused more on ensuring that products are hypervisor-independent, can scale up rapidly as the business does, and have feature parity with more traditional solutions. In other words, the vendors have focused on ensuring the solutions perform well at load.

I understand the focus on performance; it should be considered "table stakes" today, but the next area of focus needs to be on provisioning. After all, performance won't matter if the customer has to use brute force to get 15,000 phones up and running in a short time frame. This problem should be solved similarly to how it was in the telcos' environment, with integrated management solutions.

On a related and rather self-serving note, if you're interested in this topic or have something to add, please attend my UC management session at Enterprise Connect 2014 in March. This is one of many topics I hope to discuss on the panel this year.

Follow Zeus Kerravala on Twitter and Google+!
@zkerravala
Zeus Kerravala on Google+