SERVERLESS
COMPUTING
MAHDI ZELANVAR
ALIREZA HAJIMOHAMMAD EBRAHIM
2
TOPICS
Introduction
Architecture and Concepts
Challenges and Limitations
Real-World Examples and Use Cases
Future of Serverless Computing
3
INTRODUCTION
WHAT IS SERVERLESS
COMPUTING?
“Focus on your application, not on the infrastructure.”
WHY DO WE NEED
SERVERLESS?
6
ARCHITECTURE AND
CONCEPTS
7
CONCEPTS
• Event-Driven: Functions triggered by events.
• Faas (Function as a Service): on-demand functions
• Statelessness: No retained function state.
• Automatic Scaling: Resources scaled dynamically
• BaaS: backend services like databases, authentication.
• Containerization: Isolated serverless function containers
• Pay-as-You-Go: Pricing based on usage
8
EVENT DRIVEN & FASS
• Event Bridge(Event Bus)
• Messaging with SQS(Queue)
• Event Driven Execution(Lambda)
• Workflow Management(Step Functions)
• Decoupled and Scalable Services
• Monitoring and Fault Tolerance
9
STATELESS
• No Built-In Memory of Past Calls • Scalability and Isolation
• External State Management • Short-Lived Execution
10
AUTO SCALING IN SERVERLESS
• Scaling in Traditional Applications
• Cloud Providers that Offer Function Scaling
• Scaling Types:
• Scale-Out: Increases instances during high traffic.
• Scale-In: Decreases instances when traffic subsides.
11
SCALE PER REQUEST
• Reactive Scaling
• No Pre-Running Instances
• Cold Start Process
• Response Time Impact
• Instance Handling
12
CONCURRENCY VALUE AUTOSCALING
• Concurrency Handling
• Cold Start Trigger
• Queuing Support
• Asynchronous Scaling
13
METRICS-BASED AUTOSCALING
• Real-Time Metric Monitoring: Scales via resource metrics
• Slow Reaction: Slower for burst workloads
• Ideal for High-Traffic: Optimized for consistent workloads
14
BACKEND AS A SERVICE:
• Pre-Built Service
• Serverless Architecture
• Speed and Efficiency
• Automatic Scaling and Cost
Efficiency
• Security and Maintenance
15
CONTAINERIZATION
• Consistency Across
Environment
• Custom Runtimes
• Isolation and Security
• Scalability
16
PAY AS YOU GO
• No Pre-Provisioning Needed
• Scales Dynamically
• Charged Only for Usage
17
CHALLENGES AND
LIMITATIONS
MAJOR SERVERLESS PROVIDERS 18
19
COLD STARTS
20
VENDOR LOCK-IN
• Vendor Dependence • Switching Difficulty • Potential Problems
21
LIMITED CONTROL
• Less Architecture Control
• Limited Software Config
• Flexibility Challenges
22
STATELESS
Serverless functions are
stateless, treating each call
as the first.
23
DEBUGGING AND TESTING
• Challenging Debugging
• Limited Error Tracing
• Complex Testing
• Cloud Dependency
24
25
Event Injection
• Event Source Misconfiguration
• Unauthorized Access Risks
• Malicious Event Injection
• Insufficient Validation Controls
26
Insufficient logging & monitoring
Applications need monitoring to prevent charges, track issues, and secure implementations.
27
Components with known vulnerabilities
• Small, Connected Functions
• Extensive Third-Party Dependencies
• Dependency Update Risks
• Supply Chain Vulnerabilities
• New Code Component Risks
28
Broken authentication
• Cloud Accessibility •Balancing User Experience and Security
• Unwanted Access Attempts • High Target for Hackers
29
Budget exhaustion
• Autoscaling Benefits
• Pay-for-Use Model
• Denial of Wallet Attacks
• Fake Request Risks
30
Shadow APIs
• Public Cloud API Risks
• Exposure to Hackers
• Shadow API Threats
• Lack of Authentication
• Sensitive Data Exposure
31
WHERE TO USE?
WHERE TO USE
32
Real-Time Data Processing:
Scalable event-driven data handling for IoT and streaming.
Backend APIs:
Automatic scaling for responsive web and mobile experiences.
Batch processing:
Simplifies scaling for large data analysis tasks.
Image and video processing:
Efficient file handling without continuous server needs.
Chatbots:
Real-time scalability for seamless customer interactions
33
WHEN AVOID TO USE?
WHERE TO AVOID 34
Long-Running Processes:
Unsuitable for extended tasks like rendering or simulations.
Stateful Applications:
Not ideal for persistent state across function executions.
High-Latency Sensitivity:
Delays unsuitable for latency-critical applications.
Regulatory Compliance and Data Privacy:
Limited control over data storage and processing.
Tightly Coupled Systems:
Adds complexity to integrated monolithic architectures.
35
MAJOR PROVIDERS
36
REFERENCES
[1] R. Rajanikanth, Serverless Computing: Concepts, Technology, and Applications, Wiley-
IEEE Press, 2021.
[2] M. Tari, M. Ghobaei-Arani, J. Pouramini, and M. Ghorbian, "Auto-scaling mechanisms in
serverless computing: A comprehensive review," Computer Science Review, vol. 100, p.
100650, 2024.
ANY QUESTIONS?!
THANK YOU FOR
YOUR ATTENTION!