Network Design Principles: Scalability and Uptime

Network Design Fundamentals

Discuss the term scalability. What does it mean? Why is it an important network design goal? What are some challenges designers face when designing for scalability?

Scalability refers to how much growth a network design must support. It is an important goal because every company is not in a stable state; it will grow continuously. As the company grows, it requires more employees and more servers to internetwork, and still, the network should work as it used to be.

Challenges in Designing for Scalability:

• Selecting the appropriate technologies and hardware for the network design.
• Managing the traffic flow of the network effectively.
• The designer must accurately forecast the company's growth over the next two or five years.

A network design customer has a goal of 99.8 percent uptime. How much downtime will be permitted in hours per week?

Percentage downtime is $100 - 99.80 = 0.20\%$.

Total hours in a week = $7 \times 24 = 168$ hours.

Downtime in hours/week = $\frac{0.20}{100} \times 168 \text{ hours} = 0.336 \text{ hours/week}$.

Note: The calculation provided in the original text for minutes/day and seconds/hour seems inconsistent with the weekly downtime calculation based on 99.8% uptime.

Assume you are in New York City. With a 100-KB file size and 1 Gbps bandwidth between two cities, determine the propagation delay or transmission delay.

This scenario requires calculating both delays. The transmission delay is calculated based on file size and bandwidth. The propagation delay depends on the physical distance between NYC and the destination city.

Reconnaissance Attack: Learn about the tools that attackers use. Describe the tools.

Reconnaissance involves gathering information about a target network or system before launching an attack. Tools used often include:

• Nmap (Network Mapper): Used for network discovery and security auditing, often employed to map out live hosts and services.
• Whois Lookups: Used to find domain registration information, including contact details and IP address ranges.
• DNS Enumeration Tools (e.g., Fierce, DNSenum): Used to discover subdomains and associated IP addresses.

What is a 'discounting network'? Why does it pose challenges?

A discounting network (likely referring to a network with overlapping or improperly managed address spaces, sometimes related to CIDR or supernetting issues) poses challenges when a single Class A, B, or C network for which at least one route to one subnet passes through subnets of a different network. This ambiguity can lead to routing confusion and security vulnerabilities.

Why is it important to characterize a network’s logical topology and not just its physical topology? What information does a logical topology illustrate?

Characterizing the logical topology is crucial because it reveals how data actually flows, which impacts performance and security, whereas the physical topology only shows hardware connections. Missing the logical view means that:

• Scalability features might be overlooked or improperly implemented.
• Security vulnerabilities related to traffic paths may be missed.
• The possibility of missing user requirements regarding connectivity and performance increases.

A logical topology illustrates the data paths, routing relationships, and addressing schemes.

The Cooperative Association for Internet Data Analysis (CAIDA) collects information. How could CAIDA's work help a designer?

CAIDA collects extensive information about the internet's network infrastructure. This data helps a designer by providing real-world context on the logical and physical topology of the broader internet, aiding in peering decisions, performance modeling, and understanding external dependencies.

The IETF Internet Protocol Performance Measurement (IPPM) Working Group. How could this group's work help a network designer?

The IPPM group develops standards and methodologies for measuring IP performance (like latency and loss). This work helps a network designer by providing standardized metrics and tools to accurately test, validate, and ensure that their network meets specified performance objectives.

Routing Protocol Comparison

Compare and contrast distance-vector and link-state routing. If you were designing a new routing protocol, which would you use and why?

Both distance-vector and link-state protocols use algorithms to determine how to reach destination networks.

Comparison:

Protocol Selection for a New Design:

The choice depends heavily on the network design:

• If the design is simple and flat topology, distance vector might suffice due to ease of management.
• If the design is hierarchical, large, or requires fast convergence, link-state routing (or path-vector protocols) would be the preferred choice for better stability and scalability.