When setting your cloud migration strategy, it is first necessary to understand your options for moving workloads to the cloud, and when each type of transition makes the most sense.
In the table below, we’ll review common cloud migration scenarios that you can apply against your internal application set.
|Cloud migration scenario||Key characteristics|
|Lift and Shift||You’ll sometimes see this referred to as “rehosting.” This scenario involves capturing a workload running on a set of physical or virtual servers and moving these captured servers to run in a cloud environment with minimal changes to architecture and operation.|
|Extend to the cloud||This is a hybrid scenario. In this scenario, a workload that runs in a data center, hosted or virtual environment on-premises is extended to use cloud resources. A straightforward model for this is the traditional three-tier application – with database/storage, mid-tier application servers, and front-end web servers. Most often, taking advantage of cloud resources to provide a “bursting” capability by adding more front end and/or application servers to meet short term needs for additional capacity, without the need to permanently own the hardware and software environments required. It can also include virtual private clouds, as well as “local” cloud resources such as AWS Outposts or Microsoft’s Azure Stack.|
|Cloud optimized||This cloud migration strategy can be either hybrid or implemented all-in-cloud. Versus the “Extend” scenario, major portions of the application are rearchitected to take advantage of a set of cloud services. This scenario can include cloud storage for backup, DR, IaaS for scalability, cloud-deployed big data, Kubernetes/containers, Blockchain, IoT services, AI/ML offerings, and others. For a reference of the available services, check out this posting on TechRadar from February 2020 detailing the services available from Amazon Web Services (AWS).|
|Cloud Native||Completely rewrite the application for cloud deployment. Containers, PaaS environments as well as the full gamut of available cloud services are used to deliver the services of a previous application optimized for cloud delivery and cost structures.|
|Replace with SaaS||Instead of modifying or recreating an existing service to use cloud resources, replace it with an offering from the myriad already available “as a service” offerings|
But when should you choose each of these migration scenarios? And when should you just leave an existing application in place, rather than migrate it to the cloud?
For purposes of setting your cloud migration strategy, there are critical considerations that need to be assessed and understood before setting your strategy and process. The full list of assessments below doesn’t have to be completed for every application in your organization’s stack before setting your strategy – but assessments built around any applications identified in cloud migration goals should be ready before setting your cloud migration strategy.
After you complete initial application migration, it will then be time to come back and build out these critical points for other applications so that they can be prioritized and planned.
These critical considerations include:
Strategic value of the application
Is it an element of your key differentiation or go-to-market? Or is it an internal service function only? (Hint – Spend the resources on your key differentiation and customer experiences. When an application isn’t one of these, focus on reducing cost or offloading delivery wherever possible)
Application customization (if commercial software being considered for replacement by SaaS)
For applications delivering internally focused services, how heavy are the customization requirements beyond the generic off-the-shelf offerings used by other organizations? Heavy customization for your organizational structure, business environments, or physical locations, and deployments can make even a generic function like order fulfillment software difficult to replace with a SaaS type offering.
Regulatory and privacy requirements
New laws like GDPR in Europe and the California Consumer Privacy Act often require specific actions or “best practices” for safeguarding information that can make applications using these data sets more difficult to fulfill with cloud-based services or delivery. Financial and healthcare organizations also have a myriad number of these requirements to meet – and these can directly affect cloud migration).
For a mission-critical application, are there critical performance requirements for transactions and database access? These considerations can be difficult to deliver without a complete application rewrite for a cloud environment – make your first step a hybrid style deployment with key storage resources remaining in existing data centers or hosted sites (Storage performance and volume are a key part of this evaluation, and often the primary determiners here). Even to applications planned to transition to SaaS delivery, understanding performance requirements should be a part of evaluating potential suppliers.
Who are the critical stakeholders for the application? They will need to be identified, and involved in tradeoffs, decisions and compromises required to complete the task.
Discover and assess the details of your workload’s operations, resource usage, and internal or external services. Only a full understanding of this information will enable you to prioritize and understand the work needed to use one of these cloud migration options. In many cases, this discovery will identify “clusters” of services or applications that will need to be migrated together
SLAs and KPIs
Understanding application SLAs and KPIs can help hone-in on the right migration strategy. Applications with strict, high-volume or high-speed requirements tend to be mission-critical, and may be good candidates for hybrid deployments rather than all-in-cloud.
Application traffic and volume
What is the traffic for your application? If traffic is highly variable, with short peaks followed by extended periods of lighter loading, this is a good scenario for a hybrid “Extend” or “Cloud Optimized” migration as it allows your organization to take advantage of cloud resources when needed while minimizing local data center or private cloud resource requirements.
When cloud migration means “private cloud”
Migrating to a private cloud implementation (on-premises or in your hosted data center) makes most sense when workloads have consistent loads with low variability and especially with fast storage access requirements.
Infrastructure cloud zones
Will you need to have implementations in multiple cloud zones to address regional requirements? What cloud vendors can meet your requirements?
Infrastructure cloud costing and sizing
When using “Lift and Shift,” “Extend” or “ Cloud Optimized” – An assessment of your application infrastructure workload can be used to forecast your basic application needs and bursting requirements. Combine these with cloud vendor offering analysis to identify the best match from their large catalogs of available instance types and services. It’s best to optimize your workloads before capturing the data needed to select instance types and services.
Cloud vendor lock-in
Migrating to cloud environments can result in cloud vendor lock-in. If the cloud vendor’s offering is heavily customized for your use (in the case of SaaS) or uses services or APIs only available on their platform for other services, the cost of the technical transition to another platform can be the same as a complete rewrite of the application. Data can be the stickiest piece. Once your organization has terabytes to petabytes of data within a cloud environment, just copying that data to another platform could be a years-long effort. Although major cloud vendors will roll up a semi-truck to your datacenter’s back door to load your data in more quickly, don’t expect them to help you move it out of their environment the same way.
To mitigate lock-in, consider a re-architecture of the application as part of cloud migration that enables multi-cloud deployment. This enables vendor independence and have DR advantages, but can limit the services that can be used within cloud platforms (and hence the advantage gained), and also means that teams will have to master multiple cloud vendor tool environments.
Be sure to make an informed decision before deployment, understanding the tradeoffs and level of exposure and risk.
Replacing an application with a SaaS offering is often the right decision for applications that are not part of your core business differentiation or offering. If you’ve been using a Commercial-of-the-Shelf (COTS) software package for a common task like Accounting or HR that’s implemented within your data center, these are good candidates to be replaced with a SaaS offering.
But replacing these types of applications with a SaaS offering is not an immediate slam-dunk. Heavily regulated industries need to be sure that a SaaS offering meets their unique regulatory requirements (most especially Healthcare, Financial Services, and Government). Retailers need to take into account Payment Card Industry Data Security Standard (PCI DSS). And every organization should assess their risk from privacy laws that apply to their customer base or employee physical locations.
Lift and Shift is often the quickest route to migrating an application to cloud IaaS environments, but offers the least benefit. An application with multiple dependencies on existing internal services is a good candidate for VPN connections (private environment within a public cloud) that has high variability are the best candidates. Another option is to migrate a “cluster” of applications and services at one time using this scenario to an IaaS environment that includes all the services required by the combined applications.
This is the best option for initial cloud migration of a mission-critical application. Keep performance critical elements (such as storage and database servers) on-premises while making use of cloud resources when needed in a “bursting” scenario. High transaction speeds and availability characterize these scenarios. In this scenario, use of cloud resources IaaS resources is the primary change. Use of cloud storage or DR may also be in scope. Once implemented, your existing application is now a “hybrid” using both cloud and local resources.
Transition an existing code-base making use of containers, cloud storage, and other available cloud services. As noted earlier, these may include items such as cloud backup, DR, IaaS for scalability, cloud-deployed big data, Kubernetes/containers, Blockchain, IoT services, AI/ML offerings, and others. Best when enhancing an existing application to make the most of cloud resources, with a moderate level of rewriting and re-architecture.
Sometimes called “serverless computing,” use PaaS, or microservices based on Kubernetes in cloud environments or Containers as a Service (CaaS) as a platform for a completely rewritten application. Best when it’s time to transition a well-delineated, limited audience existing application to meet new needs for scalability, interoperability, or a transition to a new market approach.
We’re the critical link that can help make your cloud migration go smoothly, and keep the migrated infrastructure that underlies your applications working with optimal cost, performance, and uptime – For public, private or hybrid clouds.
- Cloud Migration Readiness: Learn about our Cloud Migration Readiness offerings that can help make the process of migrating applications to the cloud fast, efficient, and effective.
- Capture and playback workloads before deployment: Our VirtualWisdom and WorkloadWisdom products work in tandem to first optimize and capture your existing application workloads, and them play then back in the cloud to simulate operation. The result, you know it will work before throwing the switch
- AIOPs driven, application-centric infrastructure management and monitoring – anywhere: Our CloudWisdom and VirtualWisdom products combine to give you AIOps driven cloud and on-premises application-centric problem resolution, performance management/monitoring, workload automation, as well as capacity and cost management.
Click here to try our products for free.