Servers Crash Course

Table of Contents

1. Introduction & Course Overview
2. Digital Service Fundamentals
3. Introduction to Servers
4. Application Requirements & Single Points of Failure
5. Fault Tolerance & High Availability
6. Fault Tolerance Mechanisms: RAID & Clustering
7. Clustering Deep Dive
8. Physical Server Components & PC Comparison
9. Server Form Factors
10. Physical Server Management
11. Challenges of Physical Server Management
12. Introduction to Virtualization

1. Introduction & Course Overview

Welcome to our comprehensive Cloud Computing course! Before diving into AWS and cloud services, we need to build a solid foundation by understanding the underlying technologies that make cloud computing possible.

                Course Structure:
                Servers and Fault Tolerance (this session)
Virtualization concepts
Data centers
AWS Cloud Computing

            

This crash course approach ensures everyone has the necessary background knowledge before we explore AWS services. The course includes hands-on experience with labs provided within the course platform and a SandBox environment for additional practice.

2. Digital Service Fundamentals

Technology exists to serve Applications, which deliver Digital Services such as:

Websites
Desktop applications
Mobile applications

Core Layers for Digital Services:

Compute Layer: Processing power and computational resources
Storage Layer: Data storage and retrieval capabilities
Network Layer: Communication and connectivity to external systems

These three fundamental layers form the foundation upon which all digital services are built. Every application requires these layers to function properly.

3. Introduction to Servers

Server Definition: A server is essentially a powerful PC with higher hardware and software specifications designed to provide compute, storage, and network resources to run applications.

Servers are the physical infrastructure that hosts the core layers (Compute, Storage, Network) required by applications. They are specifically designed for:

Higher reliability and continuous operation
Greater processing power
Enhanced scalability
Improved fault tolerance

4. Application Requirements & Single Points of Failure

Application Requirements

When developing applications, we must consider two types of requirements:

Functional Requirements	Non-Functional Requirements
What the application does	How well it does it
Business logic, features, capabilities	Performance, latency, High Availability (HA)

Single Point of Failure (SPOF)

SPOF: A critical component whose failure will cause the entire system or application to become unavailable.

Common SPOFs in a single server include:

Power Supply
CPU
RAM
Storage (Hard disk)
Network Card

When any of these components fails, the application becomes unavailable or "down," resulting in downtime.

5. Fault Tolerance & High Availability

Fault Tolerance (FT): The ability of a system to continue operating even when failures occur within specific fault domains.

Fault Domain: The scope within which a single failure event can cause an impact (e.g., a single hard disk, a single server).

Requirements for Fault Tolerance

To implement effective fault tolerance, two essential elements are required:

Redundant Components: Having more than one of the critical components in the fault domain
Mechanism: A system to detect failures and switch to healthy redundant components

                High Availability (HA): Ensures that applications are always accessible and running, minimizing downtime through fault tolerance mechanisms.
            

6. Fault Tolerance Mechanisms: RAID & Clustering

RAID (Storage Fault Domain)

RAID: A technology that uses multiple disks (redundant components) with mechanisms to ensure data remains accessible even if one or more disks fail.

Clustering (Server Fault Domain)

Clustering: Grouping multiple servers or devices together to perform specific functions with enhanced reliability and performance.

Clustering addresses server-level failures by providing redundancy at the server level, ensuring that if one server fails, others can continue providing the service.

7. Clustering Deep Dive

Clustering serves multiple purposes and can be implemented in different ways:

1. High Availability (HA) / Failover Clustering

If the active server fails, the passive server will take over and become active.

Configuration: Active/Passive setup
Mechanism: Heartbeat monitoring between servers
Process: If the active server fails, the mechanism switches to a passive server
Data Requirement: Data replication between servers, often using shared storage

2. Load Balancing Clustering

Configuration: Active/Active setup
Purpose: Distribute workload across multiple servers to prevent overload
Mechanism: Load Balancer distributes incoming traffic based on Load Distribution Algorithms.
Benefit: Provides both performance improvement and high availability.
Note: Load balancers use listener functions that constantly check the health of backend servers before sending traffic their way.

3. High Performance Computing (HPC) Clustering

Purpose: Intensive computational tasks requiring parallel processing
Requirement: Application code designed for parallel execution
Benefit: Significantly reduced processing time for complex calculations

8. Physical Server Components & PC Comparison

Servers are distinguished from standard PCs by their enhanced components designed for reliability and performance:

Component	PC	Server
CPU Sockets	Typically 1	Multiple (2, 4, or more)
CPU Cores	2-16 cores	Many cores per socket, with Hyper-Threading
RAM Slots	2-4 slots	Many slots (16, 32, or more)
RAM Capacity	Gigabytes	Terabytes with ECC (Error Correction Code)
Storage	Single or few drives	Multiple drives with hardware RAID
Network Cards	1	Multiple for redundancy
Power Supplies	1	Multiple redundant power supplies

Key Server Advantages:

ECC Memory prevents system crashes from memory errors
Multiple power supplies ensure continuous operation
Hardware RAID provides better performance than software RAID
Higher overall reliability and uptime

9. Server Form Factors

Physical servers come in three main form factors, each optimized for different use cases:

1. Tower Servers

Appearance similar to large desktop PC cases
Stand-alone units requiring floor space
Straightforward cabling and cooling
Best for small deployments

2. Rack-Mounted Servers

Designed for horizontal installation in standard server racks
Measured in U (Units) - 1U, 2U, 4U, etc.
More space-efficient than towers for multiple servers
Complex cabling management requirements
Requires careful cooling planning due to density

3. Blade Servers

Highest density solution
Individual server "blades" slide into shared chassis
Shared infrastructure: power supplies, cooling fans, backplane
Simplified cabling and power management
Higher initial cost but better density

Factor	Tower	Rack-Mounted	Blade
Space Efficiency	Low	Medium	High
Initial Cost	Low	Medium	High
Cabling Complexity	Simple	Complex	Simplified
Cooling Requirements	Simple	Planned	High Density
Scalability	Limited	Good	Excellent

10. Physical Server Management

Managing physical servers requires different approaches depending on location and accessibility:

Local Management

KVM Switch: A device that allows a single monitor, keyboard, and mouse to control multiple servers in the same location.

KVM switches are ideal for server rooms where physical access is available and multiple servers need management from a single console.

Remote Management (Out-of-Band Management)

For servers in distant locations (e.g., data centers in other countries), remote management technologies are essential:

                IPMI-Based Technologies:
                IPMI: Intelligent Platform Management Interface
iDRAC: Dell Remote Access Controller
iLO: HP Integrated Lights-Out

            

Remote Management Capabilities

Power server on/off remotely
Configure BIOS settings
Monitor hardware health and status
Install operating systems remotely
Full server control via web browser interface
Independent network interface with its own IP address

11. Challenges of Physical Server Management & Utilization

Despite their capabilities, physical servers present significant challenges in modern IT environments:

Low Resource Utilization

Problem: Physical servers typically operate at only 15-20% average utilization of their CPU and RAM resources.

High Infrastructure Costs

Expensive server hardware
Physical space requirements
Electrical power consumption
Cooling infrastructure costs

Lack of Application Isolation

Running multiple applications directly on a single physical operating system creates two critical risks:

Surface of Failure: If one application causes problems (memory corruption, disk errors) that crash the operating system, ALL applications on that server become unavailable.

Surface of Attack: If an attacker compromises one application, they may gain access to the operating system and potentially control or attack other applications on the same server.

Migration Difficulties

Physical transportation required
Costly and time-consuming process
Impractical for long distances
Risk of hardware damage during transport

12. Introduction to Virtualization

Virtualization: A technology that allows a physical server's resources to be divided and shared among multiple isolated virtual instances.

How Virtualization Addresses Physical Server Challenges

Challenge	Virtualization Solution
Low utilization (15-20%)	Multiple virtual machines share physical resources efficiently
High infrastructure costs	Fewer physical servers needed, reducing hardware, space, power, and cooling costs
Lack of isolation	Each virtual machine is completely isolated from others
Migration difficulties	Virtual machines can be easily moved between physical servers

Types of Virtualization

Server Virtualization: Creating multiple virtual servers on physical hardware
Network Virtualization: Abstracting network resources
Storage Virtualization: Pooling storage resources

                Next Steps: This foundation in servers and virtualization prepares us for understanding data centers and cloud computing, which heavily rely on virtualization technologies to deliver scalable, efficient, and cost-effective computing services.

You now have a solid understanding of servers, fault tolerance, and the fundamental concepts that led to the development of virtualization technologies. This knowledge forms the foundation for understanding cloud computing platforms like AWS.

No contents are available in this lesson!

No lessons available !

Lesson contents locked