Medium-sized organizations face many of the same IT challenges as large enterprises, but with far fewer resources to solve them. One challenge, in particular, is how to approach the multicloud world we are progressively moving toward and operating within.
For CIOs, now is the time to get on the multicloud bus before their organizations are outpaced by competition. This post explores important rules for considering the multicloud approach.
First, it’s important to understand that ‘Multicloud’ is not a thing, it’s a strategy. Your organization is likely already using more than one cloud service, but there’s also a good chance you have not thought through that usage in a meaningful way. These four rules should jumpstart that thought process.
Rule 1: Application First, Cloud Platform Second
Your applications are core to multicloud planning and strategy. Understanding your application landscape, the requirements of each application, how and when they need to run, and external services required to interact is the starting point. Here are some specific questions to ask about your applications before considering the cloud platform best suited to them.
- What interoperability is required?
- What compliance or governance is required?
- How much data does it produce/transfer?
- What deployment tools or services are required?
- What are the usage and load patterns?
- What risks are associated to operating the particular application on one platform versus another?
- What are the unit economics I need to measure?
- What level of automation is needed?
- What is the security strategy?
These are just some of the questions to consider about your applications prior to evaluating the potential cloud platforms to support and power them.
Rule 2: Pay Attention to Cloud Economics
All clouds are not created equal. This is true both from a feature perspective, but also for economics. Running an always on, high I/O, high data transfer workload on AWS is not usually the best choice financially. For example, we’ve worked with many customers during their evaluation phase, comparing a hosted private cloud to AWS or Azure, and have shown anywhere from 20-45 percent savings to run the same application in a hosted private cloud. However, if you have workloads that are highly burstable or that only run for a few hours a day, AWS will generally be more economically viable because of the on/off capability and short term billing.
The term ‘rightsizing’ is often referred to when it comes to the economics of cloud hosting. If your virtual server options come in three sizes, but your application requires a size somewhere in between, you still end up paying for the additional resources you are not using. This is the typical scenario for most public clouds. On the other hand, you could opt for a model that allows you to select the exact amount of resources you need now, and then scale them as you need. We see this with customers using a hosted private cloud deployment model.
The net of the economics discussion is that there is no one-size fits all and that careful consideration should take place before making a selection of the right cloud platform.
Rule 3: Review Platform Feature Set Availability
What cloud platform features are vital to IT’s ability to execute on the broader organizational strategy? Once again, not all clouds are created equal in this regard. In fact, there are even nuances inside the word cloud that can blur this decision process. Infrastructure as a Service (IaaS), Platform as a Service (PaaS), Software as a Service (SaaS) are all flavors of cloud that could have the right feature availability for a given need.
If we stay focused on IaaS for a moment, we get into areas such as API availability, automation capabilities, and native feature sets that allow an IT organization to achieve a growing list of outcomes. AWS, for example, regularly releases an average of 500-700 new features per year, not including upgrades or enhancements to existing features. Approximately 720 features in 2016 alone, with an additional 400 upgrades or enhancements. On the surface this seems great for IT orgs, but what it really means is an increasing amount of complexity and a decreasing ability for the average IT department to keep up. More features generally means more cost as well.
A laundry list of features could be useful in some scenarios, but overkill in others. You have to decide on which features are important to your particular scenario, run your cost benefit analysis, determine if you have the skills to implement, and then move forward. If you’re not equipped to do this evaluation or would like a second set of eyes, our team of certified solution engineers are here to help.
Rule 4: Have a Plan to Manage Multiple Cloud Platforms
Last, but certainly not least, is the question of how you will manage the cloud platforms in your multicloud strategy. Leading CIOs are working hard to reduce complexity, not add to it, so being able to manage as many cloud services from one place is a key to success.
“Do more with less” is a theme for most IT organizations, so asking the team to manage many cloud services each with their own tools, reporting, billing, monitoring, and so on becomes quite daunting even for the most savvy of teams. This is why it’s important to seek out the ability to bring everything under one roof, so to speak. Having a single pane of glass to monitor, manage, and report from reduces complexity and ultimately gives valuable time back to focus on more strategic initiatives that are important to the business’s bottom line.
The four points above boil down to a single sentence: A multicloud strategy should enable your organization to meet specific workload or application requirements—both technically and commercially.
Updated: January 2019