When it comes to cloud migration, preparation is crucial to project success. By progressing through the following key stages you can achieve a better chance of running a smooth migration with minimum disruption.
Cloud adoption by UK companies has now neared 90%, according to the Cloud Industry Forum, and it won’t be long before all organisations are benefiting to some degree from the flexibility, efficiency and cost-savings of the cloud.
Moving past the first wave of adoption we’re seeing businesses ramp up the complexity of the workloads and applications that they’re migrating to the cloud. Perhaps this is the reason that 90% is also the proportion of companies that have reported difficulties with their cloud migration projects.
This is frustrating for IT teams when they’re deploying cloud solutions that are supposed to be reducing their burden and making life simpler. Performing a pain-free migration to the cloud is achievable, but preparation is crucial to project success.
Progressing through the following key stages offers a better chance of running a smooth migration with minimum disruption:
1. Set goals at the outset
Every organisation has different priorities when it comes to the cloud, and there’s no ‘one cloud fits all’ solution. Selecting the best options for your organisation means first understanding what you want to move, how you’ll get it to the cloud and how you’ll manage it once it’s there.
You also need to identify how migrating core data systems to the cloud will impact on your security and compliance programmes. Having a clear handle on these goals at the outset will enable you to properly scope your project.
2. Assess your on-premises
Preparing for cloud migration is a valuable opportunity to take stock of your on-premises data and applications and rank them in terms of business- criticality. This helps inform both the structure you’ll want in your cloud environment and also the order in which to migrate applications.
Ask the hard questions: does this application really need to move to the cloud or can it be decommissioned? In a cloud environment, where you pay for the resources you use, it doesn’t make economic sense to migrate legacy applications that no longer serve their purpose.
Once you have a full inventory of your environment and its workloads, you need to flag up those specific networking requirements and physical appliances that may need special care in the cloud. This ranked inventory can then be used to calculate the required cloud resources and associated costs.
Importantly, this process can also be used to classify and prioritise workloads which is invaluable in driving costs down in, for example, cloud- based disaster recovery scenarios where different workloads can be allocated different levels of protection.
3. Establish migration tech support
Many organisations take their first steps into the cloud when looking for disaster recovery solutions, enticed by the facility to replicate data continuously to a secondary location with virtually no downtime or lost data.
This is fundamentally the same as a cloud migration, except that it is planned at a convenient time, rather than prompted by an extreme event. This means that once the switch is flipped, the migration should be as smooth as a DR event.
However, most organisations will want to know that there is an expert on hand should anything go wrong, so 24/7 support should be factored into the equation.
4. Boost what you already have
Look at your on-premises environment and work out how to create synergies with the cloud. For example, VMware- users will find there’s much to be said for choosing a VMware-based cloud environment which is equipped with tools and templates specifically designed for smoothly transitioning initial workloads and templates.
It’s an opportunity to refresh the VM environment and build out a new, clean system in the cloud. This doesn’t mean you can’t transition to a cloud that differs from your on-premises environment, but it’s a factor worth taking into consideration.
5. Migration of physical workloads
Of the 90% of businesses that reported difficulty migrating to the cloud, complexity was the most commonly cited issue, and you can bet that shifting physical systems is at the root of much of that. They are often the last vestiges of legacy IT strategies and remain because they underpin business operations.
You need to determine if there is a benefit to moving them to the cloud and if so take up one of two options: virtualise the ones that can be virtualised – possibly using software options – or find a cloud provider that can support physical systems within the cloud, either on standard servers or co-located custom systems.
6. Determine info transfer approach
The approach to transferring information to the cloud will depend on the size of the dataset. In the age of virtualisation and of relatively large network pipes, seeding can often be viewed as a costly, inefficient and error prone process. However, if datasets are sufficiently large, seeding may be the best option, with your service provider providing encrypted drives from which they’ll help you manually import data into the cloud. A more innovative approach sees seeding used to jumpstart the migration process.
By seeding the cloud data centre with a point in time of your environment, you then use your standard network connection with the cloud to sync any changes before cut-over. This minimises downtime and represents the best of both worlds.
7. Check network connectivity
Your network pipe will be seeing a lot more traffic and while most organisations will find they have adequate bandwidth, it’s best to check ahead that your bandwidth will be sufficient. If your mission-critical applications demand live-streaming with zero latency you may wish to investigate direct connectivity to the cloud via VPN.
8. Consider post migration support
Your migration project is complete, now you have to manage your cloud environment and get accustomed to the variation from managing on-premises applications. The power and usability of management tools should be part of the selection criteria so that you are confident you will have ongoing visibility and the facility to monitor security, costs and performance.
Furthermore, support is a crucial part of your ongoing relationship with your cloud service provider and you need to select an option that gives you the support you need, when you need it, at the right price.
As more and more businesses take the plunge and move mission-critical systems to the cloud, we’ll see the skills and experience of in-house teams increase and the ability to handle complex migrations will rise in tandem.
Until then, IT teams charged with migration projects shouldn’t be afraid to wring as much support and advice out of cloud service providers as possible so that they can achieve a pain-free migration and start reaping the benefits that only the cloud can bring.
Why disaster recovery is a must-have
Natural disasters and data breaches have been hitting the headlines this year. While it might not happen to you on such a large scale, any sort of outage can cause IT administrators and CIOs to lie awake at night wondering if they are well protected. Disaster Recovery as a Service (DRaaS) is now a mainstream use of the cloud.
This makes a lot of sense, as it helps companies avoid having to double their infrastructure, with half of it waiting to recover from full or partial outages. But there are many more reasons to opt for a cloud-based disastery recovery (DR). You can’t afford downtime!
Whether you call it IT resilience, data protection or business continuity, the reality is that you cannot afford downtime. DR is not a luxury, it’s a necessity as it is critical to be always available for your internal and external customers. You know that if you can’t be reached by a customer, they’ll simply move to your competitor instead. Or even worse, your internal customers may turn to shadow IT or other things outside your purview, which can cause a whole lot of problems further down the line.
When you look at the news, natural disasters get the biggest headlines, but normally that’s not what’s happening in your environment. Most of the time, you suffer downtime from hardware or software fault, malicious ransomware or careless users and accidents.
The biggest challenge is that IT budgets haven’t grown bigger to accommodate DR, yet it’s fast becoming mission-critical. For every single budget cycle, the first thing that tends to get cut is DR. You hope that for just one more year nothing bad is going to happen, and no matter how many times you talk to the stakeholders about how important DR solutions are, something will cause a budget block.
However, at the end of the day it’s IT who will come under the spotlight, if an actual issue occurs.
Gartner estimates that an outage can have an average impact of $5.6K for every minute of unplanned downtime, but it’s actually quite difficult to put a precise cost on downtime. It could be anything from your ordering system, to your emails, to your presence on the web – all will have varying impacts on your business performance and when you do post-event analysis, these are the tangible effects that you can put a pound sign next to.
However, what is harder, is putting a figure against the intangible costs of downtime.
Ask yourself: “What’s the perception of your customers if they can’t reach their data or application?” If you have a hundred different applications but the one which your customer really needs is down, what’s this really telling your customer about your overall ability to support their needs? Another intangible cost of downtime is customer and internal loyalty. If your competitors are up when you’re not, your customer might decide to change his provider and the cost of retaining and/or reacquiring this customer is astronomical. Finally, let’s consider the issue of ‘confidence’.
Are your internal sales team confident they can get their orders processed? Are your customers confident that you can deliver on your promise? If that confidence goes down, customer and employee retention goes down as well. The numbers are quite scary.
Disasters can make or break an organisation, but everybody thinks that it’s not going to happen to them. We’ve talked with customers over the years and found that nearly 50% of organisations protect less than half their virtual machines with a DR plan (according to a 2018 survey conducted by Veeam).
What’s even more scary is that 85% of decision makers are not confident in their current solution’s ability to recover virtual machines (according to Veeam 2017 availability report). In the past, it would be necessary to shut down the whole data centre and spend a three-day weekend failing over and running through an extensive recovery plan.
Usually, even by the end of these three days, organisations still weren’t sure this would work during a real disaster. Then Monday would come around and it would be necessary to power on and back-up anyhow, and just hope that the solution in place was enough. It’s not unusual to have some kind of IT disruption – according to Spiceworks, 77% of organisations reported experiencing at least one outage (i.e. any interruption to normal levels of IT-related service) in the last 12 months.
Simply having a power failure in your data centre can affect your whole business. According to FEMA, nearly 40% of small businesses close after a disaster. Think about that, if you lose your customers’ context, your sales receipts, orders, even if you lose your customers’ confidence, how much will it cost to regain it, and is it even possible at that point? Your backups aren’t enough!
The second reason you need a cloud- based disaster recovery is because your back-ups aren’t enough. After a disaster, you can still reinstall applications, you can always get new hardware, you can even run your internet line. Nevertheless, your business is based around your data and, if you don’t have it, you can’t operate. Back- ups are not like IT resilience or disaster recovery.
The disadvantage of back-ups, especially when you try to recover quickly, is that they are prone to unacceptable Recovery Point Objectives (RPO). If you’re only backing up every night and something happens an hour before the back-up, an entire day of workloads will be lost.
In addition, back- ups can have painful Recovery Time Objectives (RTO). If it takes you 17 hours to recover your entire data centre from tapes, in addition to how long it has been down for, that’s a considerable amount of time. On top of that, local back-ups can be targeted by malicious software, especially some nasty new ransomware packages, which target your back-up applications, database back-up files, and will search to lock them all down.
Back-ups also need something to restore to. This means that if the power in your building is out, a back-up is not going to help. If a server catches fire, a back-up is not going to help either. It’s going to take a week to get new equipment and back it up to that.
Now, this is not to say that you don’t need back-ups: they’re huge when it comes to data protection, long term archives, and security compliance, but a full IT resilience plan means back-ups and DR. Cloud makes your job easier!
There are many reasons why some organisations don’t have a DR plan. It can be a lack of manpower, skills, space or budget. However, the cloud has come along and it’s agile, scalable, highly performing, highly adaptable and cost- effective. When you think about traditional DR and purchasing a secondary data centre, duplicate hardware, replications, licences and so on, at some point you have to take into account additional headcount to help run all of that.
When you move to a cloud-based DR solution and a cloud- based service provider, you have access to the expertise of the cloud team, which makes your job considerably easier.
Most people don’t spend every single day of their lives worrying about DR or business continuity, but providers do, so they can help you create that security. We often hear people saying that if they create a secondary data centre it’s because they don’t trust the cloud or they need to have certain security and compliance solutions in place and the cloud doesn’t provide it. The reality is that in a proper cloud environment, security and compliance are built into the platform.
In fact, the cloud can be even more secure than your on-site environment. The reality is that data should be able to flow in and out effortlessly without unexpected costs. Fears about security shouldn’t be minimised because it’s “just the recovery site”.
Customers need to be in control at every step of the process. When disaster strikes, fears and unknowns are the worst things to have to deal with, so a DR solution should be simple, well-understood, secure and available.