Cloud computing, applications, and storage for business are becoming increasingly popular. But what is cloud computing, and why might you need a guide?
In this case, a ‘cloud migration’ refers to moving some or all of a business’ apps, data, and services from on-premises hardware to cloud servers and enabling stakeholder access to these new cloud solutions.
While many other options are available, this forms the core of the actions taken in most cloud migrations.
It’s not just that cloud computing allows work-from-home arrangements. Cloud computing also eases data access, centralises security, and enables access to on-demand computing resources. This on-demand nature also lets organisation scale their cloud services, spinning up more resources when demand is high and switching them off to control costs when demand drops.
These elements improve flexibility and agility in daily operations and strategic planning initiatives.
But before you jump head-on into a cloud migration project, you should know some best practices involved in the process.
When deciding to transfer some or all of your assets and processes to the cloud, it’s essential to make sure that you know why you’re making the change.
There are many benefits to utilising cloud computing, including:
An in-depth review of any migration project’s costs and benefits will let you know where you stand before you begin. Instead of jumping on the bandwagon, perform an in-depth analysis of the types of applications you are currently using. Use this to determine whether your applications or processes offer better outcomes by remaining on your servers or when moved to a cloud solution.
An experienced managed service provider can help you determine the best way to manage your assets and space when transitioning to the cloud.
Any significant project benefits from having a central point of contact for all major decisions. This step provides a transparent chain of decision-making, authority and responsibility – regardless of the project type.
The migration architect role is authorised to plan, roll out, and review any system architecture changes in any cloud migration project.
Their responsibilities include:
One of the most significant parts of the migration architect’s role is defining any refactoring required to make the migration successful. But having a clear definition of technical authority and responsibility also helps ensure the project is successful.
One of the first decisions your cloud migration architect will need to make is to choose a level of cloud integration that makes sense for your organisation, its stakeholders, and their workflows. The two choices available are known as “shallow” and “deep” clouds.
Shallow cloud integration – sometimes known as a ‘lift and shift’, shallow cloud integration involves transferring data and applications to a cloud solution with minimal changes. File structures and pathways remain broadly consistent, with any changes to applications only made to ensure they operate as expected in the cloud environment. The environment offers no cloud-unique services, and all assets remain intact in the cloud.
Deep cloud integration – the polar opposite of a shallow cloud deployment. As the name suggests, the apps and data are integrated deeply with the cloud environment to take advantage of the cloud’s unique features. This option can be as simple as utilising auto-scaling and load-balancing capabilities. Or it could be as involved as transforming working environments from hardware-based desktops to persistent user accounts that live in the cloud, with access to extra compute resources on demand.
While it is possible to transition from one level of integration to another, it pays to know which integration level works best for your organisation at the outset.
When choosing a cloud provider, you must first choose between a private and public cloud solution. More complex hybrid cloud solutions are also available, but we’ll focus on these two leading choices for brevity.
Public cloud – using the public cloud will give you greater ease of scalability. Many small and medium-sized businesses utilise pay-as-you-go public cloud services precisely because they allow for growth. These resources are available on-demand so that any activity changes can be managed quickly and at a low cost. Platforms such as AWS and Azure are especially popular with enterprise organisations with virtually limitless scalability.
Private cloud – a private cloud solution makes sense for businesses that need more customisation options or the ease of a completely managed cloud service. For example, First Focus offers a privately-hosted Smart Cloud that delivers all the public cloud solutions – including scalability, accessibility, and more – but with even greater customisation options and higher performance at no higher cost.
The advantages of our private cloud include the following:
As an MPLS provider, we can connect your office with a point-to-point network link, providing complete control and direct access to your private cloud infrastructure for faster network performance.
These benefits make a private cloud solution ideal for companies that require tailored IT solutions or prefer the single point of access that a fully managed service provides.
It’s a well-known maxim that anything you can measure, you can improve. The same is true for cloud migration projects. Defining key performance indicators (KPIs) lets you track how your applications and services perform against your projections.
CPU and memory performance – monitoring the CPU and memory utilisation helps generate valuable insights into the computing resources an application or operation uses. If the CPU or memory usage nears 100%, this can indicate an issue with an application or inefficiency in the system.
Application performance – many cloud migration projects involve moving users away from traditional server infrastructure to Platform/Software-as-a-Service (PaaS and SaaS) solutions. These services often obfuscate the individual back-end servers and make it hard to see how your application back-end responds to user demand. Tracking the performance of applications as a separate metric makes it easier to understand how well the PaaS and SaaS solutions are performing by giving functional metrics to compare to your previous infrastructure and can flag potential issues that may not be visible using other KPIs.
Access uptime – this metric tracks the amount of time your cloud-based resources are accessible by end-users. Low uptime percentages could indicate issues with the setup or hardware involved.
Error rates – this shows the frequency of HTTP status code errors. Expressed as a percentage, this KPI lets you know about connectivity issues within your cloud network.
Customer satisfaction scores – otherwise known as CSAT, this metric interprets the user’s subjective experience with a solution as an objective outcome. A project’s CSAT scores can show how migrating to the cloud has changed workflows and outcomes while measuring user sentiment towards the new cloud-based solution.
Response times – measuring the time between receiving and completing an end user’s request can help track speed and performance issues. The most common response times measurements are Average Response Time (ART) and Peak Response Time (PRT).
Error types – tracking error types lets you gain insight into the kind of issues experienced by users as they come to grips with the new cloud deployment. An increase in one error type could indicate problems with specific apps or network elements.
Indicators of compromise – IOCs flag unusual activity that could indicate potentially malicious attacks on the network. Security analysts should constantly monitor IOCs to avoid data breaches and unauthorised access. Common IOCs include geographical irregularities and unusual outbound network traffic.
Network I/O – measuring the network input and output lets security analysts check network traffic from all monitored terminals. This KPI allows security analysts to track phenomena like unusual access metrics and flag them for review.
User audits – this KPI keeps tabs on how many users have access to the cloud server, the length of time they maintained access, and the resources they accessed during their session.
A vast range of potential KPIs is available when considering a cloud migration. Across each category, it’s essential to identify which metrics are the most important to your business – including metrics likely to be the most impacted and the levels that these impacts are considered acceptable.
Once you have determined which measurements matter to your cloud migration project, it’s time to identify the baseline levels your users currently experience. This measurement process lets you diagnose any issues encountered during the migration while also tracking improvements that come about as a result.
The first step in baselining is to measure the pre-migration performance of your applications and services. Setting a short baseline of 24 hours lets you move quickly but at the risk of obtaining a sample that does not reflect general use. A more extended baseline of 30 days provides a more representative sample but takes longer and doesn’t allow for a fast turnaround.
The second choice involves deciding between collecting aggregate average data versus the metrics resulting from peak usage or critical operations. The former provides a good general level of performance, while the latter can help inform architecture decisions for environments that frequently require high-performance levels.
Regardless of which metric you use, ensuring that the baseline selected is clearly defined is essential to avoid confusion.
There’s no doubt that the actual process of migrating data is one of the thorniest parts of any cloud migration project. Many user processes and workflows have developed around on-premises access. Off-premises solutions also need access to data and apps involved in the migration. Changing these workflows and solution access without prior planning can significantly disrupt performance.
A data migration plan will let you map your move in manageable stages with minimal disruptions.
Some methods for overcoming these issues include:
When you transfer your data, you should also avoid moving everything in one go. Attempting to move vast amounts of data and applications simultaneously can slow the process down and doesn’t allow for frequent integrity checks during and after the migration.
We recommend taking a staggered approach that moves applications one by one. This approach lets you prioritise what needs moving and decide whether you can migrate some services (such as email) to the cloud first to minimise disruption and other considerations regarding the specific applications you use.
There is an element of risk for your files during large data transfers, especially if you attempt them yourself. Although they are safe once on the cloud, your files are significantly at risk during the transfer. This danger means ensuring data gets backed up before making the switch is imperative.
If you are attempting a transfer yourself, any zipped files should be password-protected, and you should only use a non-public, secure network. For extra security and efficiency, use an ethernet connection and complete your transfer in steps instead of all at once when dealing with large amounts of data.
At First Focus, we look after all file transfers and security concerns in cloud migration projects. We save secure backups before beginning a transfer, only use secure channels to perform the migration, and stagger the migration process to ensure data integrity.
Deciding when to switch from your legacy systems to your new cloud solution depends on the complexity of the architecture and procedures involved.
Once all the apps and data have been migrated according to your migration plan, there are two approaches you can follow.
Make the switch quickly – once you’ve moved your data and apps to the cloud and checked that all connections are set up correctly, direct all traffic from the on-premises system to the new cloud stack. This process allows for a fast switchover process and suits smaller projects that you can thoroughly test in advance.
Perform incremental changes – this option is a more cautious approach that involves moving a predetermined number of users from on-premises access to the cloud solution. It lets you test things and make adjustments with minimal disruption to the broader workforce. Once you’re satisfied that the new solution is performing as expected, you can transfer the next batch of users. Continue this process until you’ve moved all your users to the cloud-based solution.
Congratulations! You’ve successfully migrated all relevant apps and services to your chosen cloud solution and spun down your legacy stacks. Unfortunately, there’s still more to be done. Now is the time to check your resource optimisation.
Most legacy on-premises solutions have resources allocated statically and don’t allow for cloud solutions’ resource-sharing options. Consequently, many of the apps and services you initially deployed in-house may not be able to take advantage of the dynamic resource allocation capacities that help the cloud ensure that users and processes have what they need when they need it.
As you move into the cloud, make sure your migration plan addresses your applications and processes resource allocation methods, including when to spin up more resources and switch them off.
The ten steps in this article cover many of the processes involved when migrating to the cloud. However, every organisation has different needs, and there’s no one-size-fits-all solution.
This uniqueness can make transitioning to the cloud seem a herculean task. However, an experienced managed cloud expert like First Focus can help streamline decision-making with tailored solutions that suit your business.
A managed service provider can also arrange end-to-end support so your assistance doesn’t stop at the cloud server. Our technical account managers understand your entire configuration and are available to help with any IT issue, backed by our professional service desk and on-site technicians.
With the right advice and support, transitioning to the cloud can be a simple step toward greater efficiency and productivity.