Guide to Cloud Migration: From On-Prem to Public Cloud / Private Cloud / Hybrid Cloud

The definition of “the cloud” has become a little, shall we say, cloudy, so before we can talk about migration to a cloud environment, also known as an Infrastructure as a Service (IaaS) provider, we’ll need to set the scene.

Cloud computing has been around for longer than you might expect. Even back in the 1960s, when people first started tossing around the idea of computer networks, the notion of accessing information from a shared space outside your personal computer was already part of the plan.

Cloud computing became a reality in 2006 with the launch of Amazon’s Elastic Compute Cloud (EC2) which allowed developers to access shared, secure computing resources and infrastructure on Amazon Web Services (AWS) via the internet, allowing them to launch products without buying their own infrastructure. Nowadays, a huge part of the internet’s core infrastructure is located on AWS.

Despite the scale and complexity of global cloud-computing models, the basic idea of this kind of utility service access has been around for a while. 

Commercial power provides us with a decent analogy. Imagine a factory that decided to provide its own power. They’d need a generator big enough for their operations and mechanics to keep the thing running properly all year.

Every time they wanted to expand their operation, they’d need to source and purchase a new generator, install it, and switch over from the old generator. This means months of work and considerable capital expenditure (CapEx). 

Not the best business plan, right? That’s why most folk outsource that work and hassle to the electric company, along with many other utility services. Viewed in this way, cloud computing is the latest in a long line of problems you can now hand off to someone else.

So, let’s settle on a definition.

What is “the Cloud?”

It’s a marketing term used to describe a series of distributed servers / compute resources available on-demand for consumption via the internet without direct active management by the user. Typically, these servers and compute resources are built in a way that supports subletting portions of compute resources to tenants or customers who will pay the resource owner based on consumption. These environments are often built with an excessive capacity to allow customers to grow rapidly.

The main components of these computing environments usually include the following.

• CPU (Central processing unit)

• RAM (Random-access memory)

• File Storage (Disk Size)

• Connectivity (Typically Internet)

• Security (Firewall)

• Third-party management and maintenance of infrastructure

So, what makes these computing environments so robust? They tend to be in data centers and offer the following advantages:

• Physical security. All the servers involved are in a secure cabinet, inside a secure cage, in a secure zone of a secure building, on a property that’s…you get the picture.

• Redundant power. Power supplies are backed up from a variety of sources, including additional generators and battery arrays.

• Redundant heating, ventilation, and air conditioning (HVAC), to ensure all equipment is kept under optimal conditions for performance.

• Diverse connectivity via multiple companies, ensuring you can still access your services via a secondary provider should your primary network go down.

Enterprises can opt to self manage their own cloud by leasing space at a colocation data center or working with a private cloud provider that would allow them to deploy their own servers for maximum control, or enterprises can utilize a “public cloud” provider such as Amazon Web Services, Microsoft Azure, or Google Cloud to avoid setup costs and active management. A combination of these two approaches is what you’d term a “hybrid cloud” approach.

What Are the Main Drivers of Cloud Migration?

A “cloud migration” is the act of shifting away from On-Premise servers and compute (customer managed at a customer site) to cloud-based servers and compute. It can also refer to migrating from one cloud service or deployment type to another, but in this post we’ll focus on shifting from On-Premise to the cloud (with a particular focus on public cloud IaaS). So, why would one do this? 

Scalability

As the hapless factory owner in our commercial-power analogy would agree, it’s hard to adjust to changes in demand when you’re providing your own utilities.

IaaS is based around the idea that even though your provider has physical limits, as a customer you’ll never be aware of them. Their abundant compute resources are available to you at the push of a button. Behind the scenes, the cloud service provider is running utilization monitors which help them stay ahead of demand and ensure they can immediately satisfy any increase in demand.

Any enterprise that has tried to scale quickly, or even allocate additional compute power to an application using an on-prem server environment will understand the value of this.

An in-house IT team tasked with this upgrade will need to source, order, and pay upfront for additional hardware. Once it has been delivered, they’ll need to schedule, implement, and troubleshoot the upgrade. This whole process can take several weeks.

Productivity

IaaS offers two significant productivity benefits to enterprises.

• Continuous optimization of applications and services. The ease with which additional compute resources can be managed means nothing ever has to run slowly if you don’t want it – saving everyone time and frustration.

• Better deployment of in-house IT staff – freed from onerous infrastructure maintenance tasks, your team can stay focused on challenges more central to your business goals.

Cost

Running servers on-prem is going to involve some unavoidable CapEx. Those costs are also inflated by the need to factor in some headroom, buying more than you need to allow for growth or periods of high demand.

However, if you’re running cloud services, your costs are purely operational (OpEx), so no heavy upfront payments should be required. You can also order just what you need, as headroom can be obtained instantly.

Uptime

The best thing about outsourcing your infrastructure to a cloud provider is the fact that they’re pretty good at it – no offense to your IT team, but your in-house usually has more than a few different roles to juggle.

Meanwhile, your dedicated cloud provider is working exclusively with high-end computer hardware installed in purpose-built data centers. As a result, they’ll deliver exceptional performance and uptime – IaaS providers usually offer SLA guarantees of anywhere between “four nines” (99.99%), “five nines” (99.999%)...heck, some of them go all the way, with 100% uptime guarantees. 

It’s hard to argue with a 100% uptime guarantee – especially if you have previous experience of significant periods of downtime, and the associated knock-on productivity effects.

Do I Need Private Cloud, Public Cloud, or Hybrid?

To answer this question fully, you’ll need to take a long, hard look at your business needs to see which service provides the most cost-effective solution.

Public cloud is a multi-tenant environment (meaning you’ll potentially be sharing servers and other resources with other businesses) that’s typically accessed via the public internet. Public cloud solutions are the easiest to set up (no physical infrastructure required at all!) and require no CapEx upfront, but provide slightly less flexibility and control than Private Cloud options, and have very unfavorable cost curves once you achieve scale.

Private cloud is a single-tenant environment where a cloud computing environment is dedicated to a single customer. This can be involve an enterprise setting up their own infrastructure in a colocation data center where some infrastructure is shared (power, cooling, connectivity access) but customer servers are self-managed, or utilizing the services of a private cloud provider that will deploy a private cloud instance on your behalf, providing maximum control and flexibility. Alternatively, a private cloud Private cloud service involves upfront CapEx to purchase and install equipment, and is typically more expensive than public cloud services at small scale. At scale, well-managed private cloud deployments can be materially cheaper than public cloud deployments, and of course certain highly specific cloud deployments must be run in a private cloud environment. See here for a guide we put together on colocation pricing for reference.

Hybrid cloud, as you might expect, involves a “mix-and-match” approach, so you can get the guaranteed performance, bandwidth, and security of a private cloud for your most sensitive applications and processes, and send less business-critical data via public cloud.

How Do I Assess My Cloud Migration Needs?

Every migration is different but completing the following tasks will give you a good foundational understanding of what you’ll need to know before you can plan your migration.

Business-unit alignment. Assuming your business has multiple sites and departments, you’ll need to map out the role IT tools and services play for each separate entity or unit. Find out how each unit uses IT to accomplish their goals.

Analyze existing servers. Review each server in as much detail as you can, to capture its existing state, configurations, and the role it plays in your infrastructure. 

For some older servers and networks, this can be trickier than it sounds – if your staff turnover is high, you might find fully functioning mystery servers, and no-one knows what they do. 

And please, please, document as you go – if the worst happens and you need to roll it all back, you’ll be very glad you did.

Determine server relevancy for the new infrastructure. If you’re slimming down, and dumping some in-house apps and services, check where the associated storage for these defunct capabilities is housed – you might be able to get away with archiving the data contained on the related server, then decommissioning it.

Stay open and aware of alternative replacement technologies, such as SaaS options. There’s no shortage of different ways to do things – before you migrate that creaky old business application, find out if there’s a lighter, cheaper, more agile alternative in SaaS form. 

For example, if you’ve been using an in-house CRM app server, it’s worth checking if there’s a Salesforce, Zoho, or SugarCRM package that might take some of the weight off.

Security first. Don’t lose sight of your security – your organization is particularly vulnerable at times like these. Modern hackers are proactive when it comes to exploiting enterprises and will look for opportunities to exploit you. Plan to protect your data both while it’s at rest and when it’s migrating, and ensure that your new cloud environment is equal to (or  preferably surpasses) your current compliance and security standards. 

What Will I Need to Plan Cloud Migration?

Once you have a good idea of the shape of your migration, you’ll need to build your plan with the following non-negotiables in place.

A clear timeline. Keep your migration project on track with clear goals and milestones.

Effective communication systems. Make sure your internal comms is up to scratch so each business unit is aware of their part in the plan, along with any impacts to their operations.

The right resources. Don’t start the migration until you have all the pieces in place – that includes IT professionals with the right availability and skillset.

The correct tools. There’s cloud migration-specific project-management tools designed to assist with this job. Evaluate them in accordance with your plans, and you’ll have a smoother transition.

Will I Need to Make Back-up Plans for the Migration?

No migration is completely risk-free – but if disaster recovery and business continuity aren’t already embedded in your network management, your cloud migration is the perfect opportunity to correct that oversight.

If you need to get up to speed with this kind of planning, there are several Lightyear articles dealing with disaster and diversity in detail. Incorporate these considerations at every stage and build in regular DR testing once the migration is complete.

At the very least, make sure your critical business data is backed up to a secondary, air-gapped location in case disaster strikes, so you be sure you won’t lose everything should there be a catastrophic failure or a security breach.

Don’t let it get you too worried, though – if you strategize effectively, there’s no reason your migration shouldn’t go smoothly. There are several companies that specialize in facilitating IaaS migrations, day in, day out. If you feel unsure about any of this, get in touch with Lightyear, and let one of our networking experts help you strategize around your cloud migration approach.