Transcript Design - Welcome | EQUELLA

Network design
Topic 2
Existing network infrastructure
Agenda
•
•
•
•
•
Network maps
Network health
Network availability
Network utilisation
Network efficiency
Starting point
• Understand the existing network structure
• Document the topology and physical structure
• Understand what the network is used for and
how it behaves
– Performance
– Bottlenecks
• Design for interoperability with the existing
network and capacity for future
Characterise the network
infrastructure
•
•
•
•
•
•
Locate major internetwork devices
Locate network segments
Develop network maps
Document naming and addressing
Document types and lengths of cables
Investigate building and environment
constraints
Develop network maps
• High level map
– Showing states, cities and campuses
– WAN connections
– WAN and LAN connections between buildings and between
campuses
• Campus maps
–
–
–
–
–
–
–
Buildings and floors, rooms and cubicles
Location of major servers
Location of routers and switches
Location of mainframes
Location of network management stations
Location of VLANs
Indication of workstations (not each one)
Logical maps,
applications and services
• Locate applications and services used by network users
–
–
–
–
–
–
–
–
Custom applications
Internal web services, intranet
Web caching servers
Email
FTP
Print and file sharing
Application servers
Database servers and data stores
• Tools such as CartoReso
– http://cartoreso.campus.ecp.fr/index.php
Logical maps
network services
•
•
•
•
•
•
•
•
Security servers – TACACS and Radius
DHCP
DNS
NAT
Print Servers
SNMP
VPN devices
RAS servers
Logical maps
Layer 3 topology
• Locate routers, links and interface names
• Locate networks and subnets
– Route summarisation
– Route aggregation
• Routing protocol information
– Hot standby router protocol groupings
– OSPF areas
– Redistribution points
– Router roles such as DR for OSPF
Logical maps
Layer 2 topology
• Layer 2 devices
– Switches and access points
• WAN and LAN technologies
– Frame, ISDN, Ethernet
• Service provider for WAN links and circuit IDs
• STP information
– Root bridge location, root ports, redundant links
• VLANs
• Trunks between switches
Modular design
Topology for:
• Core
• Distribution
• Access
Wiring and media
Wiring charts
•
•
•
•
Types of cables
Cables distances
Labelling on equipment and cables
Connections between buildings including number
of wires and type of media and distance
• Location of telecommunications room and wiring
closets
• Vertical wiring runs
• Horizontal wiring runs
Building and environmental
constraints
• Within buildings:
–
–
–
–
–
–
–
Air conditioning
Heating
Ventilation
Power
Protection from EMI
Security locks
Space for cable conduits, patch panels, equipment racks
and work areas
• Between buildings:
– Flooding, heavy traffic to jostle cables, legal right of way,
line of sight for wireless
Health of existing network
• Baseline measurements on the performance
of the existing network
– Select an appropriate timeframe
• Multiple time frames and sufficient time frames
• Don’t measure for typical performance when the
network is under abnormal load
– Select a typical time period of normal
performance
– Measure for errors, packet loss and latency
Network availability
• Gather statistics
– MTBF mean time between failure
– MTTR mean time to repair for each major
segment
– Find out the causes of recent and disruptive
periods of downtime
Network utilisation
• How much bandwidth is in use during a
specific interval?
• Measure bandwidth utilisation by protocol
– Consider broadcast traffic and unicast traffic
– Relative network utilisation usage by
protocol/total usage
– Absolute network utilisation
• Use a protocol analyser to measure traffic
Measure network accuracy
• Measure the bit error rate
– Use a BER tester
– Show interface commands to find the number of
errors on serial interfaces
• Look for CRC errors on frames on Ethernet
networks
– Show interface switchport
• Measure lost packets and response times
Network efficiency
• Maximum frame sizes ensure large amounts
of data per frame and therefore the number
of frames and round trip delays are reduced
• Large window size means multiple frames are
accepted before pausing to acknowledging
• Increase window sizes on servers and clients
• Increase MTU on router interfaces
– necessary for tunnel traffic
– extra headers may exceed MTU and fragment
Delay and response time
• Send ping packets and measure RTT round trip
time
• Measure variance for realtime applications which
do not tolerate jitter
• Measure response times for typical applications
and functions:
–
–
–
–
Checking email
Sending a file
Loading a web page
Printing
Status of major routers, switches and
firewalls
•
•
•
•
CPU utilisation - How busy the device is
How many packets have been processed
How many packets have been dropped
Status of buffers and queues
–
–
–
–
–
Show buffers
Show environment
Show interfaces
Show memory
Show processes
• SNMP can also monitor the health of devices
Tools
• Protocol analysers
– EtherPeek from WildPackets
– Wireshark®
• Network monitoring tools:
– MRGT multi router traffic grapher
• Network traffic load and performance
characteristics
– Cricket
– Periscope Network Analysis
– Netflow
Agenda
•
•
•
•
•
Network maps
Network health
Network availability
Network utilisation
Network efficiency