CCDA - en/Enterprise LAN design

From Teknologisk videncenter
< CCDA - en
Revision as of 12:41, 16 September 2013 by Heth (talk | contribs) (Created page with " =Materials= *Network design chapter 3 (Loads PDF) ==Reading plan== Unless you are preparing for the CCDA certification, you don't...")
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search

Materials

Reading plan

Unless you are preparing for the CCDA certification, you don't need to read the following subjects

  • 10BASE5 and 10BASE2
  • Token Ring
  • Repeaters
  • HUB
  • Bridges
  • 5-4-3 rule

LAN media

Enterprise campus

Best Pratice for Hierarchical Layers

Access Lag

  • Limit VLAN's to the rack closet
  • Use RPVST+ if it is necessary to use VTP
  • Configure VLAN Dynamic Trunking Protocol (DTP) to Desirable/Desirable with negotiation disabled.
  • Prune VLAN manuelly to reduce broadcasts.
  • Use VTP is transparant mode. No common VLAN database necessary.
  • Disable Trunking on access ports. More swcurity and will make Portfast work faster.
  • Consider using Routing in the Access Layer to get fast convergence and load-balancing.
A1(config-if)#<input>do show run int fa0/1</input>
interface FastEthernet0/1
 description Accessport USER
 no ip address
end

A1(config-if)#<input>switchport host</input>
switchport mode will be set to access
spanning-tree portfast will be enabled
channel group will be disabled

A1(config-if)#<input>do show run int fa0/1</input>
interface FastEthernet0/1
 description Accessport USER
 <notice>switchport mode access</notice>
 no ip address
 <notice>spanning-tree portfast</notice>
end

Distribution Layer

Distribution layer
  • Wire Speed on all ports.
  • Link redundancy.
  • Use FHRP for example HSRP or GLBP is using OSI layer 2 links to Access Switches.
  • Use OSI Lag 3 links between Distribution and Core layer for faster convergens and load balancing.
  • Use distributionsswitches to connect VLAN on more than one Access Switches.
  • Summarize Routes from Distribution to Core layer reducing Routing overhead.
  • Byg Layer 3 triangles, not squares, as shown on the diagram below.
Use triangles not squares

Core Layer

Dependent on the size of the network you could inplement a Core Layer. In Campus size network the Core Layer function is high-speed Layer 3 switching between the switch blocks.

Core Layer in Campus sized network
  • Use Layer 3 Switches in the Core Layer

Network in big buildings

The Network in big buildings er segmented i floors and/or departments.

Access and Distribution Layer example in big building

Enterprise Campus Network

A campus network connects to or more building in a limited geographical area. An Enterprise consists of two or more Campuses.

  • Each building is assigned IP address spans allowing route summarization.
  • Each campus is assigned IP address span allowing Campus Route summarization to reduce routing overhead.
    • Design a logical IP address paln for example 10.campus.building.X or similar.

Edge distribution

  • Edge distribution is the interface between a Campus and the surrounding world.
    • Other Campuses in the same Enterprise
    • VPN
    • Internet

Edge Distribution protects Campus against:

  • IP Spoofing (Edge Distribution Layer protects Core layer from announcing illegal IP addresses)
  • Unauthorized access to the Core.
  • Network Reconnaissance (Blocking trafic used by Crackers to recognder kan benyttes til at reconnoiter the network)
  • Packet sniffers - (Separate Broadcast zones and limit traffic which can be sniffed)

Medium size networks

Medium sized network consists of 200 to 1000 network units. Distribution and Corelayer are often integrated in the same units. Called Collapsed Core.

Medium sized network (Collapsed Core)

Små netværk og opkobling af mindre filialer

Mindre netværk og mindre filialer er ofte koblet op til hovedkontoret med en mindre Router. Routeren fjerner Broadcast trafik fra WAN forbindelsen.

Opkobling af mindre netværk (Remote Site LAN)

Server Farm modul

Server Farm modulet eller data centeret giver højhastigheds adgang til servere for Campuset. Servere tilsluttes Switche med 1 eller 10 Gbps Ethernet. I visse tilfælde anvendes EtherChannel

Serverfarm

OPkobling af Servere

  • Single NIC
  • Dual NIC EtherChannel
  • Content Switching (Flere servere er frontend)

Enterprise Data Center infrastructure

Enterprise Data Center

Data Centeret (DC) giver high performance/low latency lag 2 switching og supporterer servere med enkelte og dobbelte netkort. Det foretrukne design er lag 2 i access laget og lag 3 i distributionslaget. Blade Chassiss[1] med integrerede Switche er blevet en populær løsning.

Data center Aggregation Layer (Distribution Layer) aggregerer - samler trafikken - og

  • Kan Load Balance trafikken mellem flere servere
  • Kan offloade servere med SSL trafik ved at terminere krypteringen i enheder.
  • Har Firewalls til at kontrollere adgangen.
  • Har Intrusion Detection Services IDS

Campus LAN Quality of Service

Klassifikation og mærkning af QoS foregår i Access Laget og Enterprise Edge (ud mod verden) og indføre QoSpolitikker i Distributions laget.

ISL CoS
IEEE802.1p/802.1Q CoS
Klassificer og marker så tæt på afsenderen som muligt

Multicast trafik

Internet Group Management Protocol - IGMP - er en protokol som anvendes mellem hosts og den lokale Lag 3 switch. Hostene anvender IGMP stil at melde sig ind og ud af Multicast grupper - altså hvilke multicast datastrømme de ønsker at lytte til.

Når der anvendes Multicast i Campus er det vigtigt at overveje hvorledes Lag 2 Switche - Access lag - behandler Multicast. For at undgå lag 2 Switche sender multicast trafik ud af alle porte kan der anvendes CGMP eller IGMP Snooping.

references