Web Applications From Small to Large Scale Capacity
0 likes12 views
Web Applications: From Small-scale to Large-scale Applications to Accommodate Large Number of Concurrent Users من تطبيقات صغيرة إلى تطبيقات واسعة النطاق لاستيعاب عدد كبير من المستخدمين المتزامنين Case Study: MoHEGate: The Gate of the Ministry of Higher Education
Web Applications From Small to Large Scale Capacity
- 1. Web Applications From Small-scale to Large-scale Applications to Accommodate Large Number of Concurrent Users من ن ر المستخدمي من ر كبيعدد الستيعاب النطاق واسعة تطبيقات إىل ة ر صغي تطبيقات ن ر امنين المي Case Study: MoHEGate: The Gate of the Ministry of Higher Education Assist. Prof. Dr. Mohamed A. Alrshah Alasmarya Islamic University, Zliten, Libya IEEE Senior Member 01-06-2025 Slideshare
- 2. Introduction Objective: To discuss challenges of scaling web applications and modern solutions. Key topics: ● Limitations of shared hosting. ● Solving single points of failure (SPOF). ● Scaling strategies: Vertical vs. Horizontal. ● Technologies: VPS, Elastic Cloud, Load Balancing.
- 3. Starting Point – Shared Hosting What is Shared Hosting? ● Cheap, Easy to set up, Managed environment. ● Multiple apps share the same server resources (CPU, RAM, bandwidth). Limitations: ● Limited capacity: Cannot handle traffic spikes. ● Typically handles tons of concurrent users. ● No isolation: "Noisy neighbors" Performance degradation and variability. ● Single point of failure (SPOF): Server crash = downtime. ● No scalability: Manual upgrades required.
- 5. Technologies for Scaling Virtual Private Server (VPS): ● Dedicated resources within a shared physical server. ● Full root access, customizable. ● Fair Prices. Elastic Cloud (AWS/GCP/Azure): ● Auto-scaling, pay-as-you-go. ● Global CDN, managed services. ● More Expensive.
- 6. How Content Delivery Network (CDN) Works? Does the CDN always meet the requirements?
- 7. Scaling Beyond Shared Hosting Solutions to Overcome Limitations: Vertical Scaling (Upgrading VPS): ● Increase CPU/RAM of a Single Server. ● Hardware Ceiling Limit. ● Operating System ceiling Limit. ● 1 VPS => No Redundancy and Single Point of Failure (SPoF). Horizontal Scaling (Multiple Servers): ● Distribute Load across Servers. ● Key Benefit: No Theoretical, Hardware or OS Limits. ● More VPSs => High Redundancy.
- 8. What if you were forced to use a private network that only provides VPSs or bare metal servers? In many cases where: ● The systems are related to national and/or government projects ● With high legal and security requirements many barriers and limitations of the allowed technologies to use will be created. So, "as an application designer or architect" you have to CREATE EVERYTHING YOURSELF! Load-balancing: Session Persistence Health Checks Scalability High-Availability: Redundant Servers Failover Mechanisms Database Replication Network Redundancy Disaster Recovery: App & Database Backups Offsite or Remote Backups High-Reliability: Automated Monitoring & Alerts Self-Healing Systems Graceful Degradation Cashing Management: Application-Level Caching CDN (Content Delivery Network) Database Caching Cache Invalidation Strategy Security: DDoS Protection: Rate limiting, Web Application Firewall. Encryption: SSL and TLS for traffic, encrypted backups.
- 9. Case Study: MoHEGate: The Gate of the Ministry of Higher Education ● MoHEGate includes both a Web Application and a Mobile Application ● Targeted Organizations: ○ 28 Universities across Libya ○ 1 Libyan Research Authority including all research centers ○ 3 National Administrations ● Services: ○ Digital ID Cards ○ Profiles and Information Management ○ Academic and Job Promotions ○ Secondments, Scholarships, Study Leaves, Sabbatical Leaves ○ Transitions among Universities and Administrations ○ Assignments to Administrative Positions ○ Social and Job Variables (Retirements, Disabilities, Deaths) ○ Reports, Statistics, Dashboards and Graphical Info-boards ● Approximate Active Users: ○ Academic Staff ≈ 25K ○ Lecturer Assistants ≈ 6K ○ Employees ≈ 30K ○ Students ≈ 300K --------------------------------------------------------- ○ Approximate Total ≈ 361K
- 10. Case Study: Horizontal Scaling with High-Availability Architecture over Multi-VPS Important Notes: ● All tools used here are 100% free ● All Servers run Rocky Linux minimum 2x Load Balancers (Active/Passive) ● HAProxy Reverse Proxy and Load Balancing. ● Keepalived for Failover Management. 4x Web Application Servers ● Blazor Server Web App / API ● Kestrel Self-Hosting + DotNet Database Server and Live Replica: ● Keepalived for Failover Management. ● PostgreSQL with replication.
- 11. MoHEGate Demo ● Web App Statistics and Graphs https://www.mohe.edu.ly ● Mobile App – QR scan to validate identity and other uses
- 12. Thanks for Paying Attention. Any Questions?