Luiz - Fotolia
As more organizations seek the benefits and cost savings of migrating from legacy to cloud infrastructure, it is crucial to accurately determine what, how and when to migrate into the modern environment.
To ensure an effective and successful cloud migration, CIOs and their organizations should conduct an extensive discovery process, aided by automated discovery tools, to determine the right applications for cloud migration and assess the cost and operational considerations of doing so.
Getting started: Assess cloud readiness and workloads
Organizations embarking on a major cloud infrastructure migration should complete an initial assessment of migration readiness and develop a high-level business use case. Once an organization establishes a general understanding of where it stands on its overall cloud journey, an action plan will help determine the gaps so the IT team is ready to migrate workloads at scale.
Additionally, a portfolio discovery and analysis exercise is required to create a detailed enterprise-level migration plan and validate its business use case. This analysis yields detailed knowledge of the legacy environment, a realistic understanding of the interdependencies of applications and workloads in the portfolio, as well as a migration wave plan detailing the sequence and scheduling of the workload migrations.
To decide how, when and which workloads or applications to migrate, organizations need a complete and accurate data set of the infrastructure and application interdependencies. This includes compute, storage, network and application documentation and a complete mapping of the interdependencies between infrastructure, applications and workloads across the entire business environment.
Automated discovery tools do the heavy lifting
Automated discovery tools can significantly accelerate the cloud migration discovery process, thus reducing the time and effort required. These tools should find and categorize every piece of hardware and software in the legacy environment, preferably with agentless technology.
They should also provide a server instance to hardware mapping and a server instance -- including related storage and networking -- to application mapping for all hardware in the environment. This highlights all dependencies of each application on the infrastructure, other applications and IT support services like message passing buses, shared databases, VDI and VPN services.
In a dynamic IT environment, automated tools can regularly update discovery maps to reflect environment changes. In this case, repeat discovery processes, which are often necessary as workloads shift from on premises to cloud service providers over months of a typical migration, become a rerun of a discovery script instead of a major manual exercise.
Automated data collection usually takes two to four weeks, with initial results available within 48 to72 hours of tool activation. The tool should be able to assimilate upward from infrastructure to applications to services, creating a bottom-up mapping of the infrastructure and its relationship with service-based workloads. An additional benefit is if the tool can start at the top of the server stack and work down to all hardware and software components and services included in the computational stack of the application and workload.
Discovery tools are helpful, but they don't provide everything required to construct a migration plan at an enterprise level. During the automated data collection period, supplemental information must be gathered independently. This includes identifying application owners, the support team for the application, business units supported by the application and workload and the relative importance of application -- i.e. mission critical, business critical, business normal, dev, QA/test.
Discovery can become an infinite time sink, therefore it's important to remember the exercise is designed to create a data set good enough to plan the migration waves at scale. At some point, an organization must decide to move forward with the data in hand.
Incomplete data rarely causes a catastrophic failure but more commonly leads to longer migration times and more expensive transition costs. Migration waves can each take four to eight weeks depending on the complexity of the workloads and will likely overlap in time.
3 types of automated discovery tools
Automated discovery tools fall into the following three categories:
Migration-specific automated discovery and optimization tools
Migration-specific automated discovery tools are purpose-built, optimized for cloud migration projects and usually adopted as part of a migration exercise. Often, they are licensed on a project basis and are useful if only a defined subset of the entire legacy environment is moving.
They are also helpful for carveout exercises where a portion of the legacy environment needs to be duplicated or migrated out of the main IT environment. An example is a divestiture of a division of a larger organization. While these tools are typically used as part of a migration project, some organizations use a professional services provider to run the discovery process and tools effectively.
Infrastructure management and monitoring platforms
Infrastructure management and monitoring platforms have extensive automated discovery and application mapping capabilities and are usually in place prior to a cloud migration exercise. Typically licensed on a per-server instance, per-year basis, these tools focus on optimizing an existing environment on an ongoing basis versus the optimization of a migration exercise.
Commonly used as part of steady-state operations, organizations should train their IT operations team to use the tools even with an outsourced service provider relationship, as this ensures the in-house team can generate reports, monitor and interpret the data accurately.
Managed service provider (MSP) platforms
MSP-level platforms are usually only available to MSPs that provide professional services support to groups of clients moving to a hyperscale cloud provider. They focus on provider capabilities rather than the migrating client and would be too complicated and costly for a typical organization to use. Recently, there is a trend of specialized migration tools being combined with cloud management platforms, or in specific cases purchased by the commercial cloud providers, effectively becoming MSP-level migration platforms for the specific provider.
Regardless of the type of discovery tools, organizations planning an enterprise-scale cloud migration should utilize the automated discovery, analysis and assessment capabilities of these tools to reduce costs and operational risks and accelerate overall migration. Furthermore, as organizations select tools, they should ensure their IT operations team learns to use the tool or use a professional services provider to run the discovery and analysis process.