Cloud Computing: Maximizing the 80%
In my last MSP Mentor post, Does ‘cloud’ take the M out of MSP?, I outlined four strategies you could use as an MSP to counter the threat to your core business posed by global cloud providers like Amazon, Google and Microsoft. This time I’d like to take a closer look at the first of these – how to simplify the deployment of your cloud, and introduce the core features you and your customers will depend on, without unnecessary expenditure of time, energy and money.
The foundation of your cloud isn't necessarily where you add value. I think of it as an 80/20 rule. For the majority of cloud providers, about 80 percent of what you do is the same as everybody else – infrastructure as a service, based on pretty much the same hardware.
It's tricky to find a sustainable way to differentiate based on your infrastructure investment. It isn't impossible – high-end servers, datacenters, even disks can be ways of offering a premium service – but the opportunities are thinner, more niche, and require more up-front investment. There are usually more lucrative ways to offer a USP (unique selling proposition) as you go up the pyramid from hardware at the bottom to the service layer at the top, and that's something we'll look at in another article.
Before tackling that, however, there is plenty to do at the infrastructure layer, and it mostly happens at the level of your cloud management platform. There are a few things to consider. Here are some of the most important.
Open source, off-the-shelf or DIY?
Your business may have its own preferences, because of IT strategy or culture, but I would suggest that one factor should override everything: time to market. If you don't have a cloud already, off-the-shelf is the only way to go.
You can't afford to spend 6 months building your own from scratch, even if you do have the development budget, and the same applies to clouds built with open source components. You should be thinking in terms of weeks, not months, because the market share is slipping away from you now.
If you already have a cloud platform – and let's face it, you probably offer at least one virtualization option for your enterprise clients – there are still strong arguments for investigating alternatives that can run alongside and plug the gaps. The classic scenario is an MSP that offers services based on VMware, services that only a larger enterprise can afford. The maturity and feature set of the leading turnkey cloud platforms give you a way to create new tiers of cloud service, quickly and easily, and expand into other parts of the market.
Think lazy. There isn't much point running a cloud service that requires constant manual intervention, either from you or your customers. We're making that core 80 percent as easy as possible, and that means automating as much of the provisioning process, and as many of the day-to-day cloud management activities as possible.
A simple place to focus on this is the control panel for your cloud. Is it easy to customize and configure for your own unique take on the cloud? Is it easy to manage hypervisors and storage? To manage user permissions, and set up billing plans? Can you lock down functionality for different clients, or different users? Does it have a complete API?
The automation applies to customers too. Does it make it easy for them to order and configure their service? Can they automatically scale their own services too? This is often a point of contention. Moving from a fixed resource, plan-based kind of service model (with nice consistent revenues) might seem counterintuitive, but the ability to scale down as well as up is one of primary reasons people want to move workloads to the cloud. It's what your customers want, and giving it to them increases stickiness and loyalty.
This is also related to time-to-market, since your aim is to avoid development time and seek a cloud platform that has everything you need, out-of-the-box. The obvious place to begin is with the core IaaS cloud management/cloud orchestration functionality. Knowing what to look for is worthy of a series of articles in itself, but a reasonable starting point would be the features offered by the cloud providers your customers are already moving to, whether it's local cloud hosting specialists, or companies like AWS.
Cloud management platforms have matured extremely quickly in the competitive public cloud arena, and features that may once have seemed advanced – particularly in areas like failover, autoscaling, self-provisioning and billing – should really now be considered as standard.
So too should support for different cloud models. Public, private, VPS, hybrid – you should be looking for a platform that can work for different lines of business, and support different tiers of service – not one that handles "enterprise", and another for "SME", and yet another for services based on traditional dedicated servers. A successful cloud business squeezes as much value from its infrastructure investments as possible, by bringing as many services onto the platform as possible.
The increasing standardization of the core feature set has driven the inclusion, into cloud management systems, of new service areas that complement IaaS. These are features like storage, DNS, CDN, and as cloud providers bring more services together, there is exponential value in the ability to manage each of these services through the same portal.
Sometimes, this 'single pane of glass' approach is seen as a great way to cannibalize existing business in those other service areas. It isn't. It's a great way to maximize ARPU, Average Revenue Per User. You have to remember, it's a single pane of glass for your customers too: it makes all of your services more visible and accessible, and that always helps you sell more.
Both of these things should be on your cloud platform checklist: the scope of what it'll let you do, beyond simple IaaS; and the ease with which it lets you manage those additional service areas.
We'll look at this in more detail next time. Until then, I'd love to know what you think.
Kosten Metreweli is Chief Commercial Officer at OnApp.