AWS Migration Scoping – Malayan Colleges & CloudReady

Participants & Roles

• Victoria Po – Lead Development Representative (LDR) / Demand Generation; primary liaison with Sir Raymond
• Germaine – Account Manager for Malayan Colleges (visibility & continuity)
• CloudReady Technology
  • Matt – President, call lead & solution architect
  • Jessica – Sales representative
  • Denzel – Managed-services / technical support
• Malayan Colleges
  • Sir Raymond – IT lead / requester
  • Sir Lawrence – Co-lead
  • Sir Patrick ("PJ") – Database administrator
  • Sir Jose – Budget & procurement input
  • Additional internal admins for domain, security, networking

CloudReady Technology Background

• AWS-first partner since 2016 (≈ 9 years in operation)
• Presence in the Philippines & Singapore
• Core competencies:
  • Workload migration to AWS
  • Post-migration managed services & optimization
• Familiarity with educational sector requirements and traffic patterns

Current Situation & Objectives (Malayan Colleges)

• 10 – 15 on-premises servers targeted for migration
• Mix of workloads:
  • Student-facing LMS / portal (web + SQL Server database)
  • Two Domain Controllers (DC1, DC2)
  • Passbolt (self-hosted password vault)
  • SIEM / OT security servers
  • FTP / file server (mainly internal)
• Active student population ≈ $4000$ (peak grows during enrolment/exams)
• Key decision: buy new on-prem hardware vs. move to cloud during 2024 server refresh
• Existing annual CapEx for IT hardware ≈ $40{\,}000{\,}000$ PHP (5-year hardware life ⇒ $\approx 20{\,}000{\,}000\,\text{PHP/yr}$)

Principal Concerns Raised

• Database is resource-intensive (CPU, RAM, IOPS)
• Cost visibility, especially Data Transfer Out (DTO) charges
• Latency between web tier & DB if split across on-prem/cloud
• Security: VPN connections, inbound rules, disaster-recovery posture
• Team has limited experience in building entire infra on AWS despite prior single-instance use

AWS Cost Components Explained

1. Compute (EC2/RDS)
   • Price proportional to instance family + size + hours used
   • Example: $\text{r6i.large} \approx 80\,USD/\text{month}$ when run 24×7
2. Storage (EBS / RDS)
   • Example: $100\,GB \approx 10\,USD/\text{month}$ (gp3 baseline)
3. Data Transfer
   • Inbound to AWS: $0\,USD$ (always free)
   • Outbound from AWS to Internet: $\approx 120\,USD/TB$ in ap-southeast-1
   • Cross-AZ traffic inside same region incurs a small fee; within same AZ $0\,USD$
4. Managed services premiums (e.g.
   • RDS licensing for SQL Server
   • Elastic Load Balancer hours + LCUs)

Illustrative Cost Calculation

• Assumptions during call:
  • Average payload per full page load: $50\,\text{MB}$ (to be reduced via caching/compression)
  • Concurrent event: all $4000$ students log in once
• Traffic volume per burst:
  $4000\,\text{users} \times 50\,\text{MB} = 200\,\text{GB} = 0.2\,\text{TB}$
• DTO cost for that burst:
  $0.2\,\text{TB} \times 120\,\text{USD/TB} = 24\,\text{USD}$
• Large customer reference: 300 k students, 20 k concurrent → $\sim 4\,\text{TB}$ DTO/day ⇒ $\approx 25{,}000\,\text{PHP/day}$

Proposed AWS Architecture (High-Level)

• Region: ap-southeast-1 (Singapore) for lowest latency to PH
• Two-Tier Pattern
  1. Elastic Load Balancer (ALB)
  • Distributes traffic across Auto-Scaling Group (ASG)
  1. Web Tier – EC2 instances (stateless)
  • Suggested families: $m5$ / $c6$
  • ASG scales out during enrolment/exams and in to min 1-2 instances (e.g.
    00:00-06:00) for cost efficiency
  1. Data Tier – Amazon RDS for SQL Server (or self-managed EC2)
  • Memory-optimized family ($r6i$) for heavy DB workload
  • Multi-AZ deployment for HA
• Supporting Services
  • Amazon S3 + CloudFront for static assets (reduces DTO per page)
  • AWS Directory Service or self-hosted DCs in isolated subnets
  • VPN or AWS Site-to-Site VPN for hybrid connectivity
  • Security groups, NACLs, WAF, GuardDuty, etc.

Instance Sizing & Rightsizing

• Need actual utilization metrics (min/avg/max) for:
  • CPU (%)
  • RAM (GB)
  • Disk IOPS / throughput
  • Used vs. provisioned storage
• Goal: avoid over-provisioning; scale vertically/horizontally only when required

Data Transfer & Bandwidth Strategies

• Keep web & DB in same AZ to eliminate cross-AZ fees and latency (< 10 ms typical)
• Enable:
  • HTTP/2 + gzip/Brotli compression
  • Front-end & DB query caching
  • CDN (CloudFront) for global edge delivery
• Monitor DTO with AWS Cost Explorer & CloudWatch metrics

Server Inventory Requirements (Action for Malayan)

• For each of the 10-15 servers provide:
  • Hostname / function (e.g.
    LMS-WEB-01)
  • vCPU & RAM allocation
  • Average & peak CPU/RAM utilization
  • OS & version (Windows Srv, Ubuntu, etc.)
  • Storage type, total GB, used GB, IOPS profile
  • Network dependencies (ports, protocols, public/private)
  • Special licensing (SQL Server Edition, Windows CALs, security tools)
• Inventory will feed the AWS Pricing Calculator and architecture diagram

AWS Proof-of-Concept (POC) Funding Program

• Submit sizing sheet & calculator to AWS → eligibility for POC credits
• Duration: $1{-}2$ months
• Environment mirrors production but is free of charge to customer during test
• Activities during POC:
  • Pilot migration of key workloads
  • Load & failover testing
  • Cost & performance baselining

Migration & Planning Timeline (Draft)

1. Now – Aug 22
   • Malayan compiles full inventory & utilization metrics
   • CloudReady analyses inventory → preliminary sizing
2. Aug 22 10 AM (virtual)
   • Deep-dive meeting, white-boarding of final architecture
3. Post Aug 22
   • CloudReady produces AWS Calculator + diagram
   • Submit for POC credits
4. POC Window (≈ Sep – Oct)
   • Deploy sandbox, migrate sample data, execute load tests
5. Decision Point
   • Compare POC results vs. on-prem refresh quote
   • Prepare management proposal & budget approval

On-Prem vs. AWS – Comparative Points

• On-Prem
  • CapEx heavy: $40\,000\,000$ PHP allocated FY-24
  • Hardware lifecycle ≈ 5 yrs, risk of obsolescence
  • Fixed capacity, manual scaling, DR requires second site
• AWS
  • OpEx/pay-as-you-go; treat compute like a utility
  • Auto-scaling and right-sizing lower idle cost
  • Built-in multi-AZ HA & global DR options
  • DTO costs must be managed (caching, CDN, compression)

Additional Technical Considerations & Recommendations

• Use Infrastructure-as-Code (CloudFormation / Terraform) for reproducibility
• Implement AWS Well-Architected Framework pillars (Security, Reliability, Cost Optimization, Performance, Operational Excellence)
• Establish IAM least-privilege model; enable MFA for admins
• Centralized logging & SIEM integration (Passbolt, security servers)
• Plan VPN / Direct Connect if sustained hybrid traffic is high
• Update DNS (Cloudflare already in use) to route via ALB/CDN
• Disaster Recovery: consider cross-region backup (e.g.
  daily RDS snapshots to ap-southeast-2)

Action Items & Next Steps

• Malayan
  • Compile & send detailed server inventory + utilization before Aug 22
  • Define acceptance criteria for POC (performance KPIs, budget thresholds)
• CloudReady
  • Prepare draft AWS Calculator once inventory received
  • Schedule on-site/virtual whiteboard (Aug 22, morning)
  • Draft architecture & POC credit request form
• Joint
  • Confirm meeting logistics (Teams link / on-site visit)
  • Evaluate CloudReady “Lunch & Learn” event on Aug 19 for additional insights

Upcoming Meetings & Events

• Aug 12–13: Internal prep window (CloudReady)
• Aug 19: Lunch & Learn event (Laguna → BGC); invitation sent
• Aug 22 10 AM: Deep-dive architecture session (2 hrs, virtual or on-site)
• Late Aug: Wrap-up & POC submission