Cloud Migration | Step-By-Step Process
ScienceSoft renders cloud migration services since 2012.
Cloud migration: summary
Cloud migration helps businesses move on-premises workloads (enterprise applications, databases, data warehouses, virtual desktops, etc.) to a cloud infrastructure to achieve higher workload performance and reduce IT infrastructure costs.
Migration of a company’s first workload takes from 1-2 months (for small workloads) to 6+ months (for medium and large workloads). Subsequent workloads require less time to migrate due to the gained experience.
The migration costs are $5,000 – 15,000 for small workloads and $20,000 - $500,000 for medium/large workloads.
Cloud migration step-by-step
Every cloud migration roadmap is different, depending on the scope and scale of your migration project, the chosen migration strategy, the complexity of data storing and processing requirements, and more. The following general migration steps can be used as a checklist for your specific cloud migration project:
Step 1. Cloud migration conceptualization
Duration: 1 - 2 weeks
- Establish cloud migration goals (improving business app performance, reducing IT infrastructure costs, expanding market reach, etc.).
- Determine workloads (e.g., enterprise applications, databases, data warehouses, virtual desktops) to relocate to the cloud.
- Diagram the workloads’ topology and dependencies and perform the workloads' complexity analysis to define the scale of the cloud migration.
Step 2. Choosing a cloud migration strategy
Duration: 1 - 2 weeks
- Choose a cloud migration strategy (rehost, replatform, refactor, repurchase, retire, retain – or a combination of several strategies).
Rehosting (“lift-and-shift”)
- Migrating on-premises workloads to IaaS without code modifications.
- Quick and cheap migration to the cloud with limited cloud-driven capabilities (workload resilience, performance and scalability).
Replatforming (“lift, tinker and shift”)
- Migrating modified workloads to the cloud (no change in the workload architecture is needed).
- Improved performance, fault-tolerance, and scaling of the migrated workloads.
Refactoring (“rearchitecting”)
- Moving on-premises workloads to PaaS by making them cloud-native (or near native).
- Full scope of cloud capabilities (including cloud auto-scaling, serverless computing, etc.)
Repurchasing (“drop and shop”)
- Replacing an on-premises solution with a cloud software product (SaaS solution).
- Fast and low-risk cloud adoption.
- Choose a cloud deployment model (single cloud, multi-cloud or hybrid cloud).
Single cloud
- Deploying your workloads in a single multi-tenant cloud environment.
- Risks of vendor lock-in.
Multi-cloud
- Migrating your workloads or their parts to different public clouds.
- Leveraging key advantages of each cloud service provider (e.g., one provider offers higher database speed, while another one offers more advanced AI capabilities).
Hybrid cloud
- A combination of public and private (hosted and managed on-premises by the company itself) clouds.
- Placing the workloads with specific security requirements on the private cloud and enjoying the public cloud’s pricing benefits for other workloads.
Step 3. Cloud platform selection
Duration: 2 - 8 weeks
- Choose one or several cloud platforms (AWS, Azure, Google Cloud Platform, etc.).
- Validate the compatibility of the workloads under migration with the cloud and a particular cloud platform with the help of a Proof-of-Concept (PoC).
Step 4. Cloud migration planning and design
Duration: 3 - 4 weeks
- Calculate cloud computing resource requirements for your workloads to avoid buying redundant resources.
- Develop a risk management plan to anticipate and mitigate such risks as overpaying due to improper workload architecture, compromised data security during data migration, etc.
- Perform cloud migration cost assessment and ROI calculation.
- Establish KPIs for cloud migration (workloads’ response time, availability, error rate, etc.).
- Identify the amount of refactoring needed for the workloads under migration.
- Select tools for automated cloud migration.
- Design the architecture of the workloads to migrate.
- Prepare a cloud migration plan, prioritizing workloads to migrate and outlining a migration schedule.
Step 5. Migrating to the cloud
Duration: 4 - 22 weeks
Common migration activities include:
- Creating a new production environment based on the new workload architecture.
- Migrating data related to the workloads under migration. To keep the cloud database up-to-date, data migration should be done continuously throughout the migration process.
- Migrating workloads.
- Setting up cloud monitoring tools.
Step 6. Final testing and optimization of the cloud environment
Duration: 1 - 8 weeks
- Test the performance of the migrated solutions in the cloud environment.
- Conduct penetration testing of the new environment.
- Conduct regulatory compliance assessment (e.g., compliance with HIPAA, GDPR, etc.) and technical audit (to identify operational flaws) of the cloud infrastructure and the migrated workloads.
- Improve security (if needed) and optimize cloud performance.
Step 7. Switching from on-premises workloads to the cloud
Duration: at once or gradually within 2 - 4 weeks (for workloads with large amounts of inbound network traffic)
At this stage, the migration is complete, and your cloud environment goes live. Upon going live, you may either turn off the on-premises infrastructure of the migrated workloads or keep it for backup or as part of a hybrid cloud deployment.
Human resources required for cloud migration
Project manager
- Develops a detailed plan for a cloud migration project, including the budget, milestones, risk mitigation plan, and more.
- Manages the schedule of the cloud migration project.
- Assigns tasks to the cloud migration team and supervises their implementation.
- Prepares regular progress reports on the migration project state.
Cloud architect
- Develops a cloud migration strategy.
- Designs the architecture of the workloads to migrate and provides it to the technical team for implementation.
- Ensures that the workload architecture is aligned with business requirements.
- Develops cloud cost optimization techniques.
Software developer
- Modifies the workloads to migrate according to the architecture designed by the cloud architect.
- Performs debugging of cloud workload issues.
Database engineer
- Configures cloud database(s) and data warehouse(s).
- Implements data migration tools.
- Implements data backup/recovery and database performance tuning.
DevOps engineer
- Aligns development and operation environments using an infrastructure-as-code approach.
- Establishes CI/CD pipelines.
- Configures monitoring and alerting systems to proactively resolve cloud workload issues during and after migration.
IT security and compliance specialist
- Verifies that the migrated workloads meet security and compliance requirements.
- Provides security recommendations during the cloud migration process.
QA engineer
- Develops a cloud migration testing strategy and plan.
- Verifies the integrity of the migrated data.
- Tests the migrated workloads’ functionality, performance, etc.
Cloud migration sourcing models
The cloud migration process is handled in-house
Pros:
- Complete control over cloud migration.
- Full visibility into the cloud migration team productivity.
Cons:
- Specialized cloud migration resources (with experience in the chosen cloud services, legacy solution modernization, etc.) who are not profitable to hire for a one-time project may be required.
- Full managerial responsibility for the migration team.
Partial or full outsourcing of technical resources, the management is in-house
Pros:
- Quick project ramp-up.
- Cloud migration cost-efficiency due to high resources scalability up and down.
- Solid cloud migration expertise of a specialized outsourced team.
Cons:
- Challenges in the coordination of in-house and outsourced resources.
Outsourcing the full cloud migration process
Pros:
- The whole process is commissionned to a self-managed cloud migration team responsible for the migration outcome.
- Smooth and quick cloud migration due to established processes of an experienced cloud migration vendor.
Cons:
- Risks of contracting cloud migration to an underqualified vendor.
In-house cloud migration process is guided by an experienced advisory vendor
Pros:
- Access to an expert cloud migration team that provides knowledge sharing and specific tech help to your in-house migration team.
Cons:
- Building a collaboration process between the in-house team and the advisory vendor takes time and effort.
Apache NiFi
Used for
Migrating databases from an on-premises server to the cloud.
Description
- Graphical user interface for easier creation of various dataflows.
- Priority-based queuing with single/multiple prioritization schemes that guide how data is retrieved from a queue.
- Creating templates of frequently utilized dataflows to automate the data migration process.
- Possibility to relocate data to multiple clouds at a time.
Pricing
Free.
Rsync
Used for
Data migration from on-premises to the cloud.
Description
- Rsync accelerates data transfer by sending only the difference in the files across the link.
- Reduced latency for multiple files due to internal pipelining.
- Supported SSH to transfer data within a secure environment by encrypting network traffic between hosts. What is more, Rsync supports RSH and direct sockets for data transferring.
Pricing
Free.
Major cloud service providers (AWS, Azure, Google Cloud Platform) offer their own tools to migrate to their cloud platform:
AWS cloud migration tools and services
Cloud migration costs
The final migration costs vary greatly depending on the scope and scale of a cloud migration project. The migration team costs range from $5,000 – 15,000 (for a small workload), to $20,000 - $500,000 (for a medium/large workload).
Cloud migration cost factors
- The number and complexity of workloads to migrate and their migration readiness.
- Chosen migration strategy (rehosting, refactoring, etc.).
- Cloud deployment model (public cloud, multi-cloud, hybrid cloud). Note that multi-cloud and hybrid cloud deployment options are more expensive.
- The volume of data to transfer.
- Workload availability requirements.
- The volume of the workload refactoring and configuration activities.
- Licensing costs for migration automation software.
- Testing coverage.
- Cloud provider charges for services.
Post-migration cost factors
- Cloud infrastructure costs.
- Cloud infrastructure support and management costs.
- Cloud support team training (for in-house migration).
Consider professional cloud migration services
ScienceSoft has been providing cloud migration services for 9 years and can offer both advisory and practical migration help.
Cloud migration consulting
- Assessing your current IT infrastructure to determine workloads to migrate.
- Estimating ROI of the migration activities and TCO of the migrated workloads.
- Planning a сloud migration project.
- Advising on cost optimization of the migrated workloads, and more.
- Guiding and assisting an in-house team through the cloud migration process (architecture designing, planning, testing, knowledge transfer, etc.).
Cloud migration outsourcing
- Choosing a cloud migration strategy.
- Defining a cloud deployment model.
- Executing migration to the cloud.
- Providing cloud-specific knowledge transfer to your in-house team.
- Setting up monitoring tools to measure the migrated workloads’ performance, load, etc.
- Developing a cloud cost optimization strategy for the migrated workloads.
About ScienceSoft
ScienceSoft is a global IT consulting and software development company headquartered in McKinney, TX. Since 2012, we’ve been providing cloud migration services to help businesses increase their workloads’ performance and reduce IT infrastructure operational costs. Being ISO 9001 and ISO 27001 certified, we rely on a mature quality management system and guarantee cooperation with us does not pose any risks to our customers' data security.