Migrating Legacy Systems To The Cloud: Strategies And Best Practices

Assessing Your Legacy Infrastructure

The first step in migrating your legacy infrastructure to the cloud is performing a comprehensive assessment of your existing hardware, software, dependencies and pain points. This involves creating a detailed inventory of servers, storage systems, networking equipment, databases, applications, integration tools and any other components currently supporting your workloads.

Document hostname, model numbers, configurations, usage profiles, ownership details and metadata for each asset. Note down the purpose it serves, what workloads, services and components depend on it, and metrics like compute capacity, network throughput and storage space. Also track end of support and maintenance timelines for aging hardware.

Furthermore evaluate aspects like utilization, scalability limits, single points of failure, backup statuses and availability of disaster recovery infrastructure. Identify which systems are underpowered, facing capacity constraints or proving unreliable under peak loads. Take stock of technical debt accrued over years of incremental upgrades and changes.

Lastly audit all components for security protections and compliance with regulations like HIPAA, PCI DSS or GDPR relevant to your industry. Note gaps around encryption, access controls, patch levels, audit logging, vulnerability testing etc. that need remediation. Similarly note any licensing issues, proprietary platforms or outdated programming languages that increase lock-in risks.

Key Questions to Ask

• What hardware, OS, software and tools comprise our current environment?
• What applications, workloads and services run on which systems?
• How are different components interconnected and interdependent?
• What are the capacity, performance and availability parameters?
• What are the security, compliance and licensing statuses?
• What technical debt exists and what upgrades are pending?

Selecting Your Cloud Platform

Once you have assessed your existing infrastructure, next evaluate leading public cloud platforms like AWS, Microsoft Azure, Google Cloud Platform, IBM Cloud and Oracle Cloud. Compare their basic service models – IaaS provides virtualized hardware resources, PaaS adds developer platforms and tools while SaaS offers turnkey applications.

Factors to weigh include breadth and depth of services, global footprint of regions and availability zones, pricing models and TCO calculators, migration tools and partnerships, security and compliance offerings, service level agreements and support options. You want a robust, enterprise grade platform that can faithfully replicate your current environment.

Cloud Service Models

• IaaS: Virtual machines, storage, networking, OS images. Eg: EC2, S3, VPC on AWS.
• PaaS: Managed runtimes, databases, containers. Eg: Elastic Beanstalk, RDS on AWS.
• SaaS: Turnkey cloud applications. Eg: Office 365, Salesforce CRM.

Evaluate IaaS instances types matching your application profiles – compute optimized, memory intensive, storage heavy, GPU based etc. Select accompanying managed services for databases, analytics, queues, notifications, caching, monitoring etc. Account for dev, test, prod and regional deployment environments. Budget for data transfer charges.

Additionally assess platform native access control, encryption and security tools. Validate available compliance certifications match your industry. Account for costs, contractual terms and learning curves in your decision matrix.

Key Evaluation Criteria

• Breadth and depth of IaaS, PaaS and SaaS services
• Support for open source platforms and BYOL
• Maturity of migration tools and partnerships
• Regional footprint and high availability
• Self-service instance provisioning and automation
• Cost transparency and TCO optimization
• Security, access controls and compliance coverage

Preparing for Migration

Now lay the groundwork for migration by setting up cloud provider accounts, networks, storage, identities and access controls. Replicate initial pilot environments mirroring your current infrastructure footprint as close as possible.

Determine networking needs – calculate subnets, route tables, internet gateways, NAT, VPNs etc required. Benchmark network latency and throughput requirements for performance sensitive workloads. Evaluate direct connect or dedicated connections to cloud regions.

Estimate storage requirements considering database sizes, file shares volumes, object storage needs and archives. Account for growth projections, read-write patterns and availability needs. Choose appropriate storage classes/ tiers balancing performance and costs.

Define IAM roles, groups, policies and permissions to provide least privilege access control. Integrate federated identity systems using SAML/OIDC. Establish security groups, ACLs, password policies and MFA aligned to existing CIA standards.

Build comparable dev, test and staging environments in the cloud account to validate migration process before scheduling production cutover. Test networking, permissions and logins work as intended. Bake in security, compliance controls and automation upfront.

Key Preparation Steps

• Estimate compute, network and storage needs
• Procure cloud accounts and logins
• Model existing network topology
• Configure IAM access controls
• Build staging environments
• Set up monitoring and log aggregation

Migrating Data Stores

For migrating databases, key considerations include downtime, data syncing mechanism, integrity validation checks, optimizing for cloud native performance and re-platforming to managed database services like AWS RDS, DynamoDB etc.

Common Database Migration Strategies

• Backup Restore: Quick to setup but requires downtime. Risk of data loss.
• Snapshot Replication: Minimizes downtime. Ensures crash consistency.
• Heterogeneous Migration: Continuous synchronization. Allows schema transformations.
• In-place Migration: Incrementally shifts data while keeping schema intact.

For large and complex database environments, use specialist tools like AWS Database Migration Service (DMS) or commercial products like Shareplex, Quest SharePlex and Ispirer MnMTK which offer heterogeneous migration capabilities, minimal downtime and continuous data syncing.

Ensure adequate capacity planning on the target database while setting up replicas and migration processes. Standardize storage and partitioning schemes optimally for cloud scale. Validate all schema objects including tables, views, indexes, functions, procedures get ported over as intended.

After migration turn off source databases. Update connection strings and resources referencing old datastores to point to new target databases. Perform sanity checks by running test workloads including ETL jobs, reports etc.

Key Post Migration Tasks

• Freeze sync and switch over to new database
• Update connection strings and dependencies
• Tune workloads for cloud native scale
• Validate reporting, ETL and applications
• Backup migrated data stores
• Decommission old databases

Replatforming Applications

As you replatform applications to cloud infrastructure, evaluate both redeploying legacy software as-is using IaaS lift-and-shift techniques as well as re-architecting apps leveraging cloud native PaaS capabilities.

Lift-and-shift approaches like repackaging applications in virtual machines or docker containers allow quicker migrations with minimal code changes. However they carry over inefficient on-prem resource utilization patterns to the cloud.

Optimally combine IaaS for legacy runtimes with PaaS for cloud scale data, caching, queues, notifications. This “hybrid cloud” mix allows selectively modernizing application architectures in parallel tracks minimizing risk.

Evaluate refactoring legacy monoliths into containerized microservices that auto scale independently. Break down N-tier app silos into decoupled, distributed,event-driven cloud native patterns. Modernize UX and leverage serverless functions for intermittent workloads.

Budget time for testing cloud versions against on-prem deployments for feature, performance and security parity before cutover. Address gaps like encryption, logging, SSO integration etc. Instrument with cloud monitoring and cost reporting tools.

Key Application Migration Factors

• Business criticality and risk appetite
• Code complexity and test coverage
• Integration dependencies
• Data gravity and network effects
• Feature gap analysis
• Cloud optimized re-architecture paths

Optimizing Costs and Performance

As infrastructure gets cloud provisioned, apply automation and devops principles to optimize price performance. Tap into cloud economies of scale and elasticity to better align spend with usage patterns.

Right size workloads selecting instance types to best match application needs at lowest cost. Compute intensive apps may use optimized/burstable instances. Memory driven workloads can use high RAM instances. Storage heavy apps lean towards provisioned IOPS SSDs.

Implement auto scaling groups with policies triggered by metrics like CPU, requests served etc. to dynamically grow and shrink resources deployed. Similarly scale data stores using read replicas. This allows smoothing volatile demand saving costs.

Tap into cloud provider cost analytics tools like AWS Cost Explorer to visualize spend distribution. Drill down on usage hotspots for savings opportunities. Establish budgets alerts to control blowout risks. Tag resources systematically to track TCO.

Further optimize by purchasing reserved capacity or spot instances in advance, consolidating multiple workloads where possible and shutting down non-production resources when not active. Route traffic intelligently using latency based DNS and route 53 policies to reduce network delays.

Key Optimization Levers

• Right size instance types to workload profiles
• Auto scale resources via demand based policies
• Spot buy transient serverless capacity
• Visualize spend across services and dimensions
• Model and simulate costs using TCO tools

Securing Your Cloud Footprint

While migrating to public cloud shifts physical infrastructure management burden to providers, owning governance responsibilities around access controls, auditing, compliance and disaster recovery remains with the customer.

Establish least privilege permissions using IAM roles for both human and application identities. Require strong MFA controls commensurate with data sensitivity. Follow just-enough, just-in-time privilege models limiting standing access.

Enforce action logging through tools like AWS Cloudtrail to capture monitoring telemetry for audits. Filter unauthorized requests before reaching services using network ACLs and security groups.

Encrypt data in transit over wire and at rest using platform KMS tools or managed hardware security modules for keys storage. Enable transparent data encryption features available for cloud data stores.

additionally invest in third party security solutions like host intrusion detection (HIDS), DLP, NGFW, malware detection capabilities layered atop IaaS foundations.

Craft RPO/RTO defined continuity plans addressing disaster scenarios including instance, AZ and region failures. Test failover procedures between multi-region or multi-cloud backups. Maintain encrypted offsite data snapshots beyond cloud provider disaster recovery provisions.

Critical Cloud Security Focus Areas

• Identity and access management
• Data protection
• Threat protection
• Behavior analytics
• Network controls
• Business continuity