From colo to cloud: Climbing the services stack - Introduction

Over the past few years, multi-tenant datacenters (MTDCs) have been faced with a choice – scale up with larger facilities and geo-diversified portfolios, or scale out with service additions. Some providers have done both. For the scale-up crowd, the way forward is a little more apparent, or at least it's been done enough that it's considered a 'known.' Scaling out, though, raises questions about what services to add and why, how to pull it off technically, and how much to charge for the services – beyond the debate as to whether this is even a good idea. In light of all this, we've put together a series of reports highlighting various services we see in the market today: managed OS; security; storage and backups; hosting; and cloud, IaaS and PaaS. Each report focuses on a specific service and seeks to define what it is, why a provider should consider it, and how to pull it off (including pricing), along with a discussion of the pitfalls and risks associated with offering the service. The aim is to give company leaders ideas on what might fit in their organization, and perhaps what wouldn't.

In this initial report, we take a look at what we consider standard services in the MTDC space, and what services we see as moving up the stack. We also examine the debate surrounding the various services, and pull in some data from our Voice of the Enterprise surveys that will hopefully add a little more color to both sides of the conversation.

The 451 Take:
When making these decisions, there is no easy button. Any of the services up the stack from colocation will require at least some shift in thinking on the part of company leadership, and the success of these services may depend on the market where the business operates – because of the business makeup therein. With wise market evaluation and a commitment to quality service, moving up the stack to services beyond colocation can be a decent investment for a future benefit. For those providers that remain committed to the idea of not competing with their customer base, the way forward may well be partnering with those customers for an expanded portfolio and a shared benefit.

For the purposes of this discussion, we're calling powered shell and wholesale, colocation (including remote hands), and networking services 'standard' services. These are all well-established services, and datacenter providers have been running these types of businesses for years, and know how to hire people to fill positions inside the organization. None of this is to say that delivering these services is necessarily easy, but it is to say that these services are known.

It's worth noting that even these businesses have been changing over the last several years. Lots of the major wholesale providers in the US have acquired retail providers, allowing them to offer smaller deal sizes, while several retail providers have started offering wholesale services or are chasing larger deals than we would traditionally consider for such providers. Networking services continue to evolve, as well. Services like cross-connects, blended bandwidth, and even DDoS protection are fairly common among providers today, especially in the retail space. However, as interconnection-oriented services become more automated and software-defined, and delivered in an as-a-service model, we see these services acting as transition points into services up the stack.

Perhaps, as an example, after offering Direct Connect or ExpressRoute services into the major cloud providers, MTDCs may consider offering consulting services or migration services centered on those specific cloud providers, and so the movement starts. Finally, providers like Equinix continue to push the limits of the standard networking services we see offered from datacenter providers. The company's Cloud Exchange Fabric will connect all its datacenters once fully deployed, and offers a means to roll out connectivity through an interface that has an underlying automation engine.

The example of Direct Connect and ExpressRoute is a pertinent one, however, since even these services can elicit some amount of fear among providers. We recently sat in on a roundtable discussion with providers, some of whom were suggesting that, by offering easier access to the cloud, the provider was essentially shooting itself in the foot by ushering its customers out the door to the cloud (as if customers wouldn't move to the cloud if there was some amount of friction). MTDC providers have a clear role to play, even in a cloud-first world, but this highlights a facet of the debate happening in datacenter circles: Should we, or should we not, offer services beyond colocation, and what are the consequences of doing either?

The debate
On one side of the debate are providers that stick with colocation-type services common among MTDCs. These services are all well-defined, providers know how to sell them, they can be profitable, and customers understand them because they're used to hearing about them in the marketplace. Additionally, colocation providers feel like they're experts in colocation already. Essentially, there is nothing new to learn, and they can pull the business off easily because it's what they've always done. Finally, by sticking strictly to colocation, providers won't be seen as competitors to their managed service- and cloud service-provider customers. The most common argument we hear when meeting with providers is: 'We don't want to be in a situation where we're competing with our customers.'

On the other side of the debate are providers that offer services beyond just colocation. The reasons for doing this are to attract a new customer base and to upsell to existing customers, allowing more possible revenue streams from a single customer. This can be a win for the customer in both cases – perhaps customers would consider colocation, but need other services for it to make sense, and existing customers have the opportunity to have more services on a single bill – saving the customer from having to source these services from a litany of providers in the market. Finally, and perhaps most compelling, these additional services generally sell for a much higher price than just colocation, offering the provider the ability to make more money per square foot on their existing datacenter assets.

So, which side is right? Like most things, it depends. The figure below contains data from a recent Voice of the Enterprise survey, where we asked companies, 'If you looked at a particular colocation provider, and decided not to proceed with them, what was the reason?' The top answer: price. A few years ago, we asked companies that had left a provider about what drove their decision to do so. Again, the top answer was price. Simply put, customers are most interested in reducing their bill or having a low bill. With this in mind, there is a fair argument to be made that, if providers are going to focus on anything, they should be focusing on ways to ultimately lower their prices, and then focus on ensuring that their support, capacity and uptime are all what they should be.

Price is paramount

There is, perhaps, another angle to consider here. Datacenter providers should absolutely focus on efficiencies and ways to keep costs down, but if all a provider offers is colocation – with no other services to make itself sticky – and the provider exists in a market with any amount of competition, all it is left with is a price battle.

Unless a provider knows it can achieve best-in-class price and performance, it should be looking up the stack. When vendors get down to price wars, products tend to suffer from shrinking margins and budgets, and that's not good for anyone. It's fair to point out that most of the respondents above didn't select a lack of services as their reason for not choosing a vendor (and this is also the case in the previously mentioned survey regarding companies leaving a provider).

That said, we often hear that companies want a trusted adviser. They want a company that can help them down the road of digital transformation. Once providers gain that trust, these companies are able to go deeper with their customers, but this is all built on the foundation of datacenter services done well, at a price the customer finds reasonable. This coincides with other surveys we've conducted that show that MTDC customers' main priority is that the providers do 'datacenter stuff' very well. Again, from this we would also suggest that, if a datacenter provider doesn't get the datacenter part right, a customer looking for only colocation services will never be flipped to a customer looking for colocation and more.

Recommendations on moving forward
So, where does all this leave us? There are a number of datacenter providers in the market today, many offering similar services and price points. However, one key differentiator is often the range of managed services available from each provider. A recent 451 VotE survey indicates that organizations are showing a growing reliance on vendor-supplied management of hosted infrastructure, and are increasingly asking service providers to offer or bundle-in these services. Simply put, customers are looking to extract greater value from hosting and cloud service providers, and MTDC providers can play in this realm, as well.

For providers looking to move up the stack, there are four basic recommendations to consider up front. First, carefully evaluate what makes the most sense for new and existing customers. As an example, if target customers largely use Veeam on a VMware infrastructure, consider adding backup services around the Veeam product base, and then maybe consider some sort of hosted infrastructure based on VMware, perhaps even going so far as a cloud environment based on VMware. Coinciding with this, be realistic about budgets and goals. If there isn't a lot of budget to spend, or no in-house talent that could pull these services off, consider something a little lower in the stack and see how it goes within the customer base.

Remember that as we go up the stack, complexity increases – as do the cost to implement and (likely) the number of employees required to make it effective. This represents a significant deviation from the typical colocation business model, which is generally capital-intensive. Conversely, managed services are generally people-intensive, and scale much differently than datacenters do.

Be realistic about the risks you will be assuming, and make sure you fully understand those risks. As an example, going after compliance-driven organizations can sound like chasing an easy target, because we know up front that they have a mandate for some of these services; however, along with these mandates come very real penalties for failures. This is not to say that non-compliance-driven industries can be treated lightly, but it is to say that those industries may not have heavy fines associated with failures. Definitely something to consider.

Finally, do it right or don't do it at all. This is perhaps a bit of a no-brainer, but as has been discussed, running a good datacenter business is how providers gain trust, which can turn into more business down the road.