Hybrid Cloud vs Multicloud: Solving the Complexity Problem

Migration to cloud is a continuously unfolding journey. I compare it to a marathon, not a sprint. By now, most enterprises have at least one public and one private cloud and are beginning to realize the many benefits of cloud computing: operational efficiencies, cost savings, agility and improved productivity. Not surprisingly, they’re hungry for more.

We are now at a crucial turning point. A decade after cloud adoption began, 80% of enterprise workloads are yet to be migrated – many mission-critical, security-dependent workloads are still to come. The real enterprise value of cloud has yet to be realized.

To accelerate and simplify application modernization, companies are evolving toward a hybrid, multicloud approach that uses a mix of public and private across a multitude of cloud platforms and vendors. It’s definitely complex, but also full of promise.

It could take another decade or more for enterprise workloads to be fully migrated to cloud. That’s the reality. As you continue the journey, how can you ensure effective, efficient communication across a complex multicloud environment – and what are the best practices and strategies that can take you from complexity to simplicity?

First, let’s define hybrid cloud, multicloud and hybrid multicloud

Hybrid cloud: Hybrid cloud is a mix of public and on-prem private cloud that is orchestrated to run a single task. An “encrypted highway” enables data sharing of mission-critical data between two on-prem and public cloud using standardized interfaces. This allows organizations to deploy secure and mission-critical workloads on-prem, while using public cloud for new workloads, and to build and test new applications.

Multicloud: Wikipedia defines multicloud as “the use of multiple cloud computing and storage services in a single heterogeneous architecture.” Unlike hybrid cloud, a multicloud environment always uses more than one public cloud. I like to differentiate between multiple clouds and multicloud.  The first is what too many have – islands of cloud technology, with no clear strategy or governance that determines which cloud, for which workload, or why.

Picture1True multicloud is the thoughtful use of cloud resources so that systems work together in a way that maximizes value, minimizes cost, and is managed proactively. 70% of enterprises claim they will be implementing a formal multicloud strategy by the end of 2019.  Why? Organizations choose public cloud to avoid vendor lock-in while gaining access to the best features and latest functionality offered by public cloud providers such as Amazon Web Services (AWS), IBM, Microsoft Azure, etc.

Hybrid Multicloud: Fuelled by open source technology with its built-in data portability and interoperability, hybrid multicloud lets customers fully leverage the data services in the cloud: containers, VMware, public cloud, private cloud, services, etc. The ability to connect public cloud to on prem or move workloads to a different public cloud lets you quickly realize the benefits of “build once, deploy anywhere.”

The situation today is a little like the islands of automation that were created in the early 1980s when minicomputers were introduced. One department would buy Digital, another HP and a third one IBM. Each had its own separate network protocols and ways to communicate. When personal computers (PCs) were added, they didn’t communicate with any of them at first. While it was possible to build connections between them, it was always easier if the connections were considered upfront.

So, how do you go about realizing ‘build once, deploy anywhere’ simplicity – and avoid the complexity and islands of disparate systems? 

Solving for complexity and cost in a multicloud world

Regardless of which cloud environment you adopt, success lies in being able to holistically manage your cloud resources as if they were in one location. That’s one of the main challenges that IT organizations are facing today. The ongoing maturation of cloud computing platforms — as well as the constant movement of both workloads and data — means complexity is virtually guaranteed.

I believe this is the reality you must manage and plan for. Trying to manage resources in one cloud without considering the rest of the environment will result in wasted capacity and budget. It is estimated that business is wasting millions on unused cloud services.

What’s needed to achieve business goals is deep visibility and insight into your hybrid, multicloud environment. We need to solve the complexity problem now, before it morphs into operational failures.

Tools can help (but there is no magic solution)

Before cloud, if a problem developed, standardized escalation procedures were followed, either on site or at an outsourced data centre. Businesses had full control of the service levels they set. It is more complex in the world of multicloud. Each of the cloud providers – whether IaaS, PaaS or SaaS – have different benefits, service levels, reporting processes and escalation responses. The benefit of standardization of the service is lower cost for you, but the flip side is you must fit their model, not the other way around. So, your team needs to know: which cloud, and which container? Is the restart automated (yet)? Is it a router issue in your network, or with the cloud provider, or with an application inside the cloud?

You can see how difficult it is stay on top of an effective management function internally. As our clients go deeper into this multicloud world, they see the need for a management layer – “cloud ops” – that provides consistency and cost control in operations. Tools like Multicloud Manager and services like IBM “AppOps as a Service” tie everything together into an integrated operations centre (SOC) and help clients use the various cloud providers to maximum benefit. It acts as a level 1.5 help desk and manages the operations and escalation process so that IT – and business – have a common view of all clouds.

Although the strategic use of tools and engaging a trusted third-party services provider can help you manage this environment, there is no magic tool; indeed, too many tools can make a complex cloud environment even more difficult to manage.

To optimize cloud use, start with planning

As you migrate more workloads to the cloud, you must decide which workload should go on which cloud and how to track usage. Here are the first few steps that will provide you with the key metrics needed to make informed decisions.

  1. Develop a clear strategy for your workloads. Think through which of your workloads fit best in which cloud: public, private, hybrid or multicloud. Which workloads are more data intensive, which more compute intensive? Which are more static, which will scale more? Which cloud providers offer the right mix of cloud services? Are you using the specific features of your IaaS / PaaS providers, or are you keeping your applications open? That’s first.
  2. Know your costs. Build an economic model and business case to track the ROI on cloud adoption and workload move. Track what you are spending – is it what you expected, or more? Are you managing costs associated with moving workloads to the cloud?
  3. Define your key governance metrics. What do you want to track? Availability, response time, and cost are obvious. What about time to create new environments, respond to new requests, or the ease of managing incidents? Are you tracking against the service levels the vendor committed to, or against your own corporate targets?
  4. Refine regularly. The cloud is constantly evolving. You will need to think about how to respond to new capabilities and monitor which suppliers are exceeding your expectations. That can help you choose where new workloads go and whether it’s time to move existing ones. When you are confident that you can move a workload from one cloud to another seamlessly, then you’ve achieved true multicloud management.


Unlock new value – without lock-in

In the minicomputer era, the islands of technology gradually became better integrated as standards emerged: TCP/IP as the network protocol replaced proprietary solutions; Unix variants replaced different server operating systems, and so on. In the new cloud era, while open standards like Linux and Kubernetes exist, there are also customizations that can turn into the kind of lock-in that dominated the previous era. The question always needs to be, is the lock-in worth the value?

We can all agree that moving mission-critical enterprise workloads onto the cloud will unlock new value for your business. But how? By developing a strategy and an economic model, you can make the informed decisions that will optimize your workloads in a multicloud architecture. From there, decide on the tools (or services) that will help you achieve your objectives.

Only then will you know the changes you’ll needed in how you run your IT shop. But that’s a topic for my next blog. Stay tuned!


Read the blog Hybrid Cloud vs Multi-Cloud: What’s the Difference?

Read A Field Guide to Multicloud Management

Read The State of the Cloud Report


Published by

Mark Dymond, P.Eng

With over 20 years of consulting experience I've worked with clients in the banking, retail, travel and government sectors. Based in Toronto, Canada, currently I lead the Cloud consulting team for IBM Canada. I've led successful project and program teams in Canada, the US, and Europe of more than 200 people. Thoughts, comments and mistakes are most definitely my own!

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s