Network Planning

Here's some information -- in the form of "checklists" -- on Network Implementation Design Analysis. It is taken from Burke, Network Management, 2004, p 45

Geographical Distribution of networks: what sorts of long-haul links do you need? What sort of scale are you considering?

  1. Single office

    1. Subnets

    2. LAN

  2. Department (many offices)

    1. Subnets

    2. LAN

  3. Division (many departments)

    1. LAN

    2. WAN connections between buildings? Between cities?

  4. Organization (many divisions)

    1. Local LAN/WAN

    2. National WAN

    3. Global WAN

Subnet organization (including the subnet-v-switching question)

  1. How many will you need:

    1. bridges/switches

    2. routers

  2. Ethernet wiring

    1. cabling

    2. 10 mbps v 100mbps

    3. IP addresses: quantity, static v DHCP

  3. Wireless

    1. number of wireless hubs

    2. authentication, privacy, and security considerations


LANs (eg local Ethernets)

  1. How many of them

  2. Domain names needed (most likely for web presence)

  3. DNS configuration

  4. IP address space: size, use of private+NAT

  5. Subnets / how many, subnet masks

  6. switched Ethernet v routers

  7. Other LAN technologies (eg token ring)?


WAN

  1. How sites connect

  2. PSTN, X.25, SONET, ATM, Frame Relay, etc


Bandwidth requirements

  1. Video needs

  2. Audio needs

  3. Data needs


Service Level Agreements

  1. bandwidth constantly available

  2. peak bandwidth

  3. bandwidth available on demand


Security

  1. firewall configuration

  2. proxy servers

  3. authentication issues

  4. network intrusion detection

  5. virus/malware monitoring



Apply this sort of planning to:

 
The following is taken from Table 3.2 from Burke, Network Management, 2004: this describes some of the data we want. Note that some of it is not easily available through SNMP.

Reliability:
Faults
Availability:    MTBF

Performance (response time)
Throughput
Data packet throughput
Voice ordered packet throughput
Video bandwidth, ...
Utilization
Resource use
Policies

Redundancy

User support


The above table is about how we measure network performance; it's a checklist for what to look at.  \
Example: approach this from the perspective of a medium-sized office  with an existing network: what need attention?
     
        Bandwidth           
                Internal and External data
        Service Level Agreement: is it sufficient?
        Security
        Known problems
        Policies
                   prioritization
                   upgrades
                   QoS
         
        Data               
            collect good baseline data

Note that Burke's table 3.2 does NOT have any per-service entries!
Some software services:

 
Burke  Table 3.3: ISO Management categories. OSI/ISO defined five:
  1. Performance
  2. Faults
  3. Configuration
  4. Security
  5. Accounting
     
Problem (with ISO generally): if you need a new category, where do you start? An ISO RFC??? That said, most things can be shoehorned into one of the above categories.
         
Performance management:
    keep tabs on network saturation, queue use, application uptime          
    maintain data, set notification thresholds, run simulations, project growth
     
Fault management: 
    Physical connectivity:
        Ping, SNMP, etc
    Application connectivity: harder
         
Config management
    choice of LAN, ISP
    number of switches, switches v routers
    managing IP addrs (eg with DHCP)
    keeping track of nodes (inventory, both manual & automated)
     
Security Management
    password policies
    best practices
    know all services
    encryption
    firewalls
    audits
    intrusion detection
     
Accounting
    At my house, who's used the most bandwidth recently?
    billing: do you do that?