• Developed by Cisco Systems in 1996
• The value of information in the cache was a secondary discovery
– Initially designed as a switching path
• NetFlow is now the primary network accounting technology in the industry
• Answers questions regarding IP traffic: who, what, where, when, and how
• NetFlow version 9 an IETF standard

• What we needs
– application performance
– application-based accounting
– network security
– Network behavior, application recognition
• ‘debug ip packet’ in router?
• IP Sniffing in shared LAN (or using switch to do so)
• Port Span in switch (how about port span in router?)
• Circuit Sniffing
• Netflow
• What we prefer in backbone:
– Embeded
– Fixed length partial packet export
– Real-time filtered packet export

• Network Monitoring
• Network planning
• Security Analysis
• Application Monitoring
• User Monitoring
• Traffic Engineering
• Peering Agreement
• Usage-base Billing
• Destination sensitive billing

Defined by seven unique keys:
1. Source IP address
2. Destination IP address
3. Source port
4. Destination port
5. Layer 3 protocol
6. TOS byte (DSCP)
7. Input interface (ifIndex)
A Flow is Unidirectional!
Exported Data

1. Create and update flows in NetFlow Cache
2. Expiration
3. Aggregation?
4. Export Version
5. Transport Protocol

SrcIf SrcIPadd DstIf DstIPadd Protocol TOS Flgs Pkts SrcPort SrcMsk SrcAS DstPort DstMsk DstAS NextHop Bytes/Pkt Active Idle
Fa1/0 173.100.21.2 Fa0/0 10.0.227.12 11 80 10 11000 00A2 /24 5 00A2 /24 15 10.0.23.2 1528 1745 4
Fa1/0 173.100.3.2
Fa0/0 10.0.227.12 6
Fa1/0 173.100.20.2 Fa0/0 10.0.227.12 11
Fa1/0 173.100.6.2
Fa0/0 10.0.227.12 6
40
80
40
0
10
0
2491 15
10000 00A1
2210 19
/26
/24
/30
196
180
180
15
00A1
19
/24
/24
/24
15
15
15
10.0.23.2 740
10.0.23.2 1428
10.0.23.2 1040
41.5
1145.5
24.5
1
3
14 step1 step2
• Inactive timer expired (15 sec is default)
• Active timer expired (30 min (1800 sec) is default)
•
NetFlow cache is full (oldest flows are expired)
• RST or FIN TCP Flag
SrcIf SrcIPadd DstIf DstIPadd Protocol TOS Flgs Pkts SrcPort SrcMsk SrcAS DstPort DstMsk DstAS NextHop Bytes/Pkt Active Idle
Fa1/0 173.100.21.2 Fa0/0 10.0.227.12 11 80 10 11000 00A2 /24 5 00A2 /24 15 10.0.23.2 1528 1800 4 step3 step4
Non-Aggregated Flows
– export
Version 5 or 9 step5
Export
Packet
Payload
(flows) e.g. Protocol-Port Aggregation Scheme becomes
Protocol Pkts SrcPort DstPort Bytes/Pkt
11 11000 00A2 00A2 1528
Aggregated Flows – export Version 8 or 9

Pre-
Processing
Packet
Sampling
Filtering
Features
And
Services
IP
Multicast
MPLS
IPv6
Post
Processing
Aggregation schemes
Non-key fields lookup
Export

Enable NetFlow
Traffic
PE
Core Network
(IP, MPLS)
Export Packets
• Approximately 1500 bytes
•
Typically contain 20-50 flow records
• Sent more frequently if traffic increases on NetFlow-enabled interfaces
UDP NetFlow
Export
Packets
Collector (Solaris, HP-UX, or Linux)
Application:
Performance
Billing
Security

• Inbound traffic only (with some exceptions)
• Unidirectional flow
• Accounts for both transit traffic and traffic destined for the router
• Works with Cisco Express Forwarding (CEF) or fast switching
• Almost supported on all interfaces and Cisco IOS
Software platforms
• Provides the sub-interface information in the flow records
• 6500/7600 enables Netflow on all interfaces by default

1400/
1600/
1700
2500/
2600
3600
3700
4500/
4700
AS5300/
5800
7200/
7500/
Catalyst
5000/6500/
7600
4500
Catalyst
ESR
10000
GSR
12000

Usage
Time of Day
QoS
• Packet Count
• Byte Count
• Start sysUpTime
• End sysUpTime
• Input ifIndex
• Output ifIndex
• Type of Service
• TCP Flags
• Protocol
• Source IP Address
• Destination IP Address
• Source TCP/UDP Port
• Destination TCP/UDP
Port
• Next Hop Address
• Source AS Number
• Dest. AS Number
• Source Prefix Mask
• Dest. Prefix Mask
From/to
Application
Routing and
Peering
Blue – Key Field (7)
Red - Lookup Field (5)
Black- Value Field (6)

• ip flow-export version <version> [origin-as | peer-as | bgp-nexthop]
– e.g. ip flow-export version 5
• ip flow-export destination <address> <port>
– e.g. ip flow-export destination 10.0.0.1 65001
• ip flow export source <interface>
– default is interface with best route to collector. Recommendation: configure loopback interface.
• ip flow-aggregation cache <name of aggregation scheme>
– select the aggregation cache
• ip flow-cache timeout inactive <seconds>
– sets the seconds an inactive flow will remain in the cache before expiration.
15 seconds is default
• ip flow-cache timeout active <mintues>
– sets the minutes an active flow will remain in the cache bvefore expiration.
30 minutes is default
• ip flow-cache entries <number>
– sets the maximum number of flow entries in the cache. The default varies dependent on platform.

• show ip cache [verbose] flow
– shows Netflow statistics
• show cache flow aggregation <name of aggregation scheme>
– shows netflow statistics for the configured aggregation scheme
• show ip flow export
– shows export statistics
• clear ip cache flow
– clears netflow statistics
• clear ip flow stats
– clears export statistics

IP packet size distribution (2175M total packets):
1-32 64 96 128 160 192 224 256 288 320 352 384 416 448 480
.001 .440 .139 .014 .008 .000 .000 .000 .000 .000 .000 .000 .011 .000 .000
512 544 576 1024 1536 2048 2560 3072 3584 4096 4608
.000 .000 .000 .002 .377 .000 .000 .000 .000 .000 .000
IP Flow Switching Cache, 4456704 bytes
550 active, 64986 inactive, 509378135 added
3145787062 ager polls, 0 flow alloc failures
Active flows timeout in 30 minutes
Inactive flows timeout in 15 seconds
Protocol Total Flows Packets Bytes Packets Active(Sec) Idle(Sec)
-------Flows /Sec /Flow /Pkt /Sec /Flow /Flow
TCP-WWW 10431912 2.4 10 181 25.7 7.1 20.4
TCP-SMTP 773843 0.1 6 98 1.1 8.3 16.7
…….
Total: 509377507 118.5 4 567 506.4 1.7 15.9
SrcIf SrcIPaddress DstIPaddress Pr SrcP DstP Pkts
Te7/3 219.245.101.77 202.205.5.3 tcp 1444 1203 1
Te7/3 84.97.234.47 202.204.192.18 udp 7692 2881 1
Te7/3 222.81.87.163 202.205.3.203 tcp 1172

Router> sh ip flow export
Flow export v5 is enabled for main cache
Exporting flows to 192.168.1.2 (2055) 192.168.2.3 (2054)
Exporting using source interface Loopback0
Version 5 flow records, origin-as
998016649 flows exported in 33267252 udp datagrams
0 flows failed due to lack of export packet
0 export packets were sent up to process level
0 export packets were dropped due to no fib
0 export packets were dropped due to adjacency issues
0 export packets were dropped due to fragmentation failures
0 export packets were dropped due to encapsulation fixup failures
0 export packets were dropped enqueuing for the RP
0 export packets were dropped due to IPC rate limiting

• Adds NetFlow switching support for:
 Cisco Catalyst 5000 Series Switches with an RSM
 Cisco Catalyst 5000 Series Switches with an MSFC
• Uses MultiLayer Switching (MLS) or CEF with Cisco
Catalyst 6000 Series Switches with SUP2
• IP unicast only
 No multicast or IPX, even if MLS can do all three
•
MLS cache is the equivalent of the NetFlow cache

• Router-based aggregation
• Enables router to summarize NetFlow data
• Reduces NetFlow Export data volume
• Decreases NetFlow Export bandwidth requirements
• Currently 11 aggregation schemes
 Five original schemes
 Six new schemes with the TOS byte field
• Several aggregations can be enabled simultaneously

 Fixed formats (versions 1, 5, 7, and 8) are not flexible and adaptable
 Cisco needed to build a new version each time a customer wanted to export new fields
 When new versions are created, partners need to reengineer to support the new export format

• Version 9 is an export format
• Still a push model
• Sent the template regularly (configurable)
• Independent of the underlying protocol, it is ready for any reliable protocol (ie: TCP,
SCTP)
• Advantage: we can add new technologies and data types quickly
• E.g. MPLS, IPv6, BGP Next Hop, Multicast

• NetFlow Version 9 Export format is template based. Version 9 record format consists of a packet header followed by at least one or more template or data FlowSets. A template FlowSet
(collection of one or more template) provides a description of the fields that will be present in future data FlowSets.
Templates provide an extensible design to the record format, a feature that should allow future enhancements to NetFlow services without requiring concurrent changes to the basic flow-record format.
• template composed of type and length
• flow records composed of template ID and value
• sent the template regularly (configurable), because of UDP

To support technologies such as
MPLS or Multicast, this export format can be leveraged to easily insert new fields
Flows from
Interface A
Flows from
Interface B
Header
(version,
# packets, sequence #,
Source ID )
Template FlowSet
Template
Record
Template ID #1
Template
Record
Template ID #2
(specific Field types and lengths)
(specific Field types and lengths)
Data FlowSet
FlowSet ID #1
Data Record Data Record
(Field values) (Field values)
Data FlowSet
FlowSet ID #2
Data Record
(Field values)
Option
Template
FlowSet
Template ID
(specific
Field types and lengths)
Option Data
FlowSet
FlowSet ID
Option
Data
Record
Option
Data
Record
(Field values)
(Field values)
• Matching ID #s is the way to associate Template to the Data Records
• The Header follows the same format as prior NetFlow versions so Collectors will be backward compatible
• Each Data Record represents one flow
• If exported flows have the same fields then they can be contained in the same Template Record e.g. unicast traffic can be combined with multicast records
• If exported flows have different fields then they can’t be contained in the same Template Record e.g. BGP nexthop can’t be combined with MPLS Aware NetFlow records

test(config)# ip flow-export version ?
1
5 Configuring Version 9 export
9 test(config)# ip flow-export version 9
Export versions available for standard NetFlow flows
Configuring Version 9 export for an aggregation scheme test(config)# ip flow-aggregation cache as test(config-flow-cache)# enabled test(config-flow-cache)# export ?
destination Specify the Destination IP address version configure aggregation cache export version test(config-flow-cache)# export version ?
8 Version 8 export format
9 Version 9 export format test(config-flow-cache)# export version 9
Export versions available for aggregated NetFlow flows

• IPFIX is an effort to:
– Define the notion of a "standard IP flow"
– Devise data encoding for IP flows
– Consider the notion of IP flow information export based upon packet sampling
– Identify and address any security privacy concerns affecting flow data
– Specify the transport mapping for carrying IP flow information(IETF approved congestion-aware transport protocol)
– Netflow version 9 has been selected as a basis for the IPFIX protocol

• PSAMP agreed to use IPFIX(Netflow version9) for export
• PSAMP is an effort to:
– specify a set of selection operations by which packets are sampled
– describe protocols by which information on sampled packets is reported to applicatons
• http://www.ietf.org/html.charters/psampcharter.html
• Note: Netflow is already using some sampling mechanisms

Access
Distribution Core Distribution Access
• Attack Mitigation
• User (IP) monitoring
• Application monitoring
• Billing
• Chargeback
• AS Peer
Monitoring
• Traffic
Engineering
• Traffic
Analysis
• Billing
• Chargeback
• AS Peer
Monitoring
• Attack Mitigation
• User (IP) monitoring
• Application monitoring
• Aggregation
Schemes (v8)
• “show ip cache flow” command
• Arbor Networks
• NetFlow
MPLS Egress
Accounting
• BGP Next-hop
(v9)
• Multicast
NetFlow (v9)
• MPLS Aware
NetFlow (v9)
• BGP Next-hop
(v9)
• Sampled
NetFlow
• NetFlow
MPLS Egress
Accounting
• BGP Nexthop (v9)
• Multicast
NetFlow (v9)
• Aggregation
Schemes (v8)
• “show ip cache flow” command
• Arbor Networks

 Flat-rate billing does not necessarily scale
 Competitive pricing models can be created with usagebased billing
 Usage-based billing considerations
 Time of day
 Within or outside of the network
 Application
 Distance-based
 Quality of Service (QoS) / Class of Service (CoS)
 Bandwidth usage
 Transit or peer
 Data transferred
 Traffic class

 Who are my top N talkers, and what percentage of traffic do they represent?
 How many users are on the network at a given time?
 When will upgrades affect the least number of users?
 How long do users spend connected to the network?
 Where Internet sites do they use?
 What is a typical pattern of usage between sites?
 Are users staying within an acceptable usage policy (AUP)?
 Alarm DOS attacks like smurf, fraggle, and SYN flood
 Will watch for these attack, regardless of source / destination

Service Provider
•
Peering arrangements
• Network Planning
• Traffic Engineering
• Accounting and billing
• Security Monitoring
Enterprise
• Internet access monitoring (protocol distribution, where traffic is going/coming)
• User Monitoring
• Application Monitoring
• Charge Back billing for departments
• Security Monitoring

Account per network (rather that per IP addresses)
Example: charge the department for the cost of the Internet link
Finance Internet
HR R&D

Account per BGP AS, to
Review Peering Agreements
ISP

Public Routers 1, 2, 3 Month of
September —Outbound Traffic
8%
6%
4%
2% 1% 1%
1%
1%
1%
1%
1%
1%
1%
8%
10%
20%
32%
Uunet
Digex
Erols
BBN
AT&T
AMU
C&W
JHU
PACBell Internet Service
RCN
OARnet
SURAnet
Compuserve
OL
ABSNET
WebTV
WEC

IP Fields
MPLS Fields
Traditional
NetFlow Fields
Source and destination IP address
Input and output sub-interfaces
Transport layer protocol
Source and destination application port numbers
8 bit IP Type of Service (ToS)
TCP Flags (accumulation from all packets in the flow)
Up to three incoming MPLS labels with experimental (EXP) bits and end-of-stack (S) bit
Position of each of the three labels
Type of the top label
IP address associated with the top label
Number of packets
Number of bytes (count either IP or MPLS header / payload)
Time-stamps of first and last packets in the flow

Traditional NetFlow for IP to MPLS traffic
MPLS Aware NetFlow
(version 9)
MPLS
Egress MPLS NetFlow Accounting for MPLS to IP traffic
IP
IP
PE
P PE
Traffic Flow
Egress MPLS NetFlow Accounting
•
IP information only
•
Ideal for billing
• Current availability: Cisco IOS Software Releases 12.0(10)ST and 12.1(5)T
MPLS Aware NetFlow (version 9)
•
Exports up to three MPLS labels, and IP packet information
• Ideal for Traffic Engineering
• Will be available in Cisco IOS Software Releases 12.0(24)S, 12.2S, and 12.3

• Origin-AS
Specifies that export statistics include the origin autonomous system (AS) for the source and destination
• Peer-AS
Specifies that export statistics include the peer
AS for the source and destination
3600-4(config)# ip flow-export version 5 ?
origin-as record origin AS peer-as record peer AS
<cr>
3600-4(config)#

AS 101 AS 102 AS 103
NetFlow enabled
AS 104
Configuring Peer-AS
• Source AS = AS 103
•
Destination AS = AS 105
Router(config)#ip flow-export version 5 peer-as
AS 105
AS 106

AS 101
AS 102 AS 103
NetFlow enabled
AS 104
AS 105
Router(config)#ip flow-export version 5 origin-as
Configuring Origin-AS
• Source AS = AS 101
• Destination AS = AS 106
AS 106

• Supported only in version 9 export
• For traffic engineering/analysis and possible billing applications
• Fields that are exported include all those found in version 5 export
• Will be supported in Cisco IOS Software
Releases 12.0(26)S, 12.2S, and 12.3

• Supported only in version 9 export
• For traffic engineering/analysis (traffic matrix) and possible billing applications. "What is the Next hop
IP address of my BGP traffic?"
• exported fields include all version 5 fields, including
IP next hop
• Adds 16 bytes to each Netflow flow record (goes from 64 bytes to 80 bytes), while CPU increase is negligible
• Edge to Edge traffic matrix for engineering/analysis and possible billing applications
• Supported in Cisco IOS Software releases 12.0(26)S,
12.2(18)S, and 12.3(1)

pamela(config)# ip flow-export version ?
1
5
9
Configuring Version 9 export pamela(config)# ip flow-export version 9 .
Configuring Version 9 export with BGP next-hop pamela(config)# ip flow-export version 9 ?
bgp-nexthop record BGP NextHop origin-as record origin AS peer-as record peer AS
<cr> pamela(config)# ip flow-export version 9 bgp-nexthop

Three types of NetFlow implementations for
Multicast traffic:
1. Traditional NetFlow
2. Multicast NetFlow Ingress
3. Multicast NetFlow Egress

(S, G) - (10.0.0.2, 224.10.10.100)
Traditional NetFlow configuration
Interface Ethernet 0 ip route-cache flow
NetFlow
Collector server ip flow-export version 9 ip flow-export destination 127.0.0.1 9995
10.0.0.2
Eth 0
Eth 1
Eth 3
127.0.0.1
Eth 2
Flow Record Created in NetFlow Cache
SrcIf SrcIPadd DstIf DstIPadd Protocol TOS Flgs SrcPort SrcMsk DstPort DstMsk NextHop Bytes Packets Active Idle
Eth 0 10.0.0.2
Null 224.10.10.100 11 80 10 00A2 /24 00A2 /24 23100 21 1745 4
•
There is only one flow per NetFlow configured input interface
• The 7 Key fields that define a unique flow are marked in red
• Destination interface is marked as “Null”
•
Bytes and Packets are the incoming values

(S, G) - (10.0.0.2, 224.10.10.100)
Multicast NetFlow Ingress configuration
NetFlow
Collector server
Interface Ethernet 0 ip multicast netflow ingress
10.0.0.2
Eth 0
Eth 1 ip flow-export version 9 ip flow-export destination 127.0.0.1 9995
Eth 3
127.0.0.1
Eth 2
Flow Record Created in NetFlow Cache
SrcIf SrcIPadd DstIf DstIPadd Protocol TOS Flgs SrcPort SrcMsk DstPort DstMsk NextHop Bytes Packets Active Idle
Eth 0 10.0.0.2
Null 224.10.10.100 11 80 10 00A2 /24 00A2 /24 69300 63 1745 4
• There is only one flow per NetFlow configured input interface
•
The 7 Key fields that define a unique flow are marked in red
• Destination interface is marked as “Null”
•
Bytes and Packets are the outgoing values

Multicast NetFlow Egress configuration
Interface Ethernet 1 ip multicast netflow egress
Interface Ethernet 2 ip multicast netflow egress
Interface Ethernet 3 ip multicast netflow egress
(S, G) - (10.0.0.2, 224.10.10.100)
10.0.0.2
Eth 0
Eth 1
Eth 3
NetFlow
Collector server
127.0.0.1
ip flow-export version 9 ip flow-export destination 127.0.0.1 9995 Eth 2
Flow Records Created in NetFlow Cache
SrcIf SrcIPadd DstIf DstIPadd Protocol TOS Flgs SrcPort SrcMsk DstPort DstMsk NextHop Bytes Packets Active Idle
Eth 0 10.0.0.2
Eth 1 224.10.10.100 11 80 10 00A2 /24 00A2 /24 23100 21 1745 4
Eth 0 10.0.0.2
Eth 2 224.10.10.100 11
Eth 0 10.0.0.2
Eth 3 224.10.10.100 11
80
80
10
10
00A2
00A2
/24
/24
00A2
00A2
/24
/24
23100 21
23100 21
1745
1745
4
4
• There is one flow per Multicast NetFlow Egress configured output interface
• One of the 7 Key fields that define a unique flow has changed from Source Interface to Destination Interface
• Bytes and Packets are the outgoing values

 Supported via NetFlow version 9 export format
 Availability
 Cisco IOS Software Releases 12.0(27)S, 12.2S, and 12.3
 Not supported in 120000
 Performance: Ingress vs. Egress
 Multicast NetFlow Ingress and traditional NetFlow will have similar performance numbers
 Multicast NetFlow Egress will have performance impact that is proportional to the number of interfaces on which it is enabled (include input interface)
 Cisco Catalyst 6500/7600 Series Switches
 Do not currently support the tracking of multicast traffic via
NetFlow due to current ASIC limitation
 Will have this support in a future Supervisor

• Suddenly highly-increased overall traffic in the network
• Higher CPU and memory utilization of network devices
• Unexpectedly large amount of traffic generated by individual hosts
• Increased number of accounting records generated
• Multiple accounting records with abnormal content, like one packet per flow record (e.g. TCP SYN flood)
• A changed mix of traffic applications, e.g. a sudden increase of "unknown" applications
• An increase of certain traffic types and messages, e.g. TCP resets or ICMP messages
• An increasing number of ACL violations

1. Cost Saver
• “sh ip cache flow” command to find top volume flows
• Identify source of attack
• Write access-list to block
• Monitor via “show ip cache flow” & “Null” entry in DestIf field to show that it is blocked
• Prefixport aggregation can be configured, while “sh ip cache flow aggregation prefixport” is used
2. Most Effective
• Arbor Networks leverages NetFlow to provide a quicker response and more sophisticated solution

ToS bits
Precedence bits
DS5 DS4 DS3 DS2 DS1 DS0 ECN ECN
128 64 32 16 8 4 2 1
DiffServ field
AKA
IP DSCP markings
Early Congestion
Notification
(ECN) bits

TOS byte
DS5 DS4 DS3 DS2 DS1 DS0 ECN ECN
128 64 32 16 8 4 2 1
Precedence bits
1 1 1
0
0
1
0
0
1
1
1
0
0
1
1
0
0
0
1
0
1
0
1
0 x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x
Delay, Throughput, and Reliability bits
Delay bit x x x x x x
0
1 x x x x
Throughput bit x x x x x x x x x x x x x x x x x x
0
1 x x x x x x
Reliability bit
0
1 x x x x x x x x x x x x x x x x
Decimal Precedence Function
224 7 Network Control (link layer keepalives)
192
160
128
96
64
32
0
0
16
0
8
6
5
4
3
2
1
0
Internetwork Control (Routing Protocols)
CRITIC/ECP (Express Forwarding)
Flash Override (Class 4)
Flash (Class 3)
Immediate (Class 2)
Priority (Class 1)
Routine (Best effort)
Delay - normal
Delay - low
Throughput - normal
Throughtput - high
0
4
Reliability - normal
Reliability - high
Early Congestion Notification (ECN) bits
ECN-capable Transport (ECT) bit x x x x x x x x x x x x x x x x x x x x x x x x
0
0
1
1
Congestion Experienced (CE) bit
0
1
0
1
0
1
2
3
Not ECN-capable
Endpoints of transport protocol ECN-capable
Endpoints of transport protocol ECN-capable
Congestion experienced

7200-3-netflow# show ip cache verbose flow
SrcIf SrcIPaddress DstIf DstIPaddress Pr TOS Flgs Pkts
Port Msk AS Port Msk AS NextHop B/Pk Active
SR6/0 210.210.210.2 PO1/0 200.200.200.2 FF 00 10 21K
0000 /0 0 0000 /0 0 0.0.0.0 1496 665.4
SR6/0 210.210.210.2 PO1/0 200.200.200.2 06 C0 00 21K
0000 /0 0 0000 /0 0 0.0.0.0 1496 666.0
7200-3-netflow# show ip cache verbose flow
SrcIf SrcIPaddress DstIf DstIPaddress Pr TOS Flgs Pkts
Port Msk AS Port Msk AS NextHop B/Pk Active
Et1/1 52.52.52.1 Fd4/0 42.42.42.1 01 55 10 3748
0000 /8 50 0000 /8 40 202.120.130.2 28 17.8
Et1/2 52.52.52.1 Fd4/0 42.42.42.1 01 CC 10 3568
0000 /8 50 0000 /8 40 202.120.130.2 28 17.8
Et1/2 10.1.3.2 Fd4/0 42.42.42.1 01 C0 10 1124
0000 /0 0 0000 /8 40 202.120.130.2 28 17.8
Hex
C0
CC
Decimal
192
204
Reliability - high
Binary
55 85 0101 0101 of transport protocol ECN-capable
1100 0000
1100 1100
Precedence 2 - Immediate (Class 2), Delay - low, Reliability - high, Endpoints
Precedence 6 - Internetwork Control (Routing Protocols)
Precedence 6 - Internetwork Control (Routing Protocols), Throughput - high,

 Deterministic
 Original type
 Cisco 12000 Series Internet Routers
 Cisco Catalyst 6500 Series Switches – Release 12.1(13)E
 Random (recommended per statistical principles)
 Cisco IOS Software Releases 12.0(26)S, 12.2S, and 12.3
 Cisco 2500, 2600, 3600, 7200, and 7500 Series Routers
 Cisco 12000 Series Internet Routers
 Time-based
 Cisco Catalyst 6500 Series Switches – Release 12.1(13)E
 Trajectory (Hash-based)
 in development

• GSR 12xxx (IOS Version: 12.0(31)S2:
R1(config)# ip flow-sampling-mode packet-interval 256
R1(config-if)# ip route-cache flow sampled input
R1(config-if)# ip route-cache flow sampled output bj2-bgw(config)#ip flow-sampling-mode packet-interval ?
<10-16382> Specify the packet interval at which to sample
• 7609: (12.2(18)SXD6)
R1(config)# mls flow ip source
R1(config)# mls nde sender version 5
R1(config)# mls sampling time-based 64 // 64:1
R1(config-if)# ip route-cache flow
R1(config-if)# mls netflow sampling

• Export is centrally via the supervisor and MSFC, each line card has its own hardware NetFlow cache and forwarding table, i.e. distributed platform

• Engine 0 – software support
• Engine 1 – software support
• Engine 2 – supported in ASICs, but lower priority so beware if running many other features
• Engine 3 – version 5 support in software, version 8 support in ASIC
• Engine 4 – not supported
• Engine 4+ – supported in ASICs

Engine
0
1
2
3
4
4+
Supported
Full NetFlow Sampled
NetFlow
Not supported

• Deterministic
– Cisco C6500/7600 Series switches(12.1(13)E)
– Cisco 12000 series internet routers (12.0(11)S and
12.0(14)ST)
• Random (select packet to export per statistical principles)
– Cisco IOS Software Releases 12.0(26)S, 12.2S(18), and
12.3(1)T
– Cisco 800, 1700, 1800, 2600, 2800, 3600, 3700, 3800, 7200, and 7500 series routers
• Time-based
– Cisco C6500/7600 series Random and Time based sampling
12.1(13)E

• Additional CPU utilization
Number of Active Flows
10,000
Additional CPU Utilization
<4%
45,000 <12%
65,000 <16%
• NetFlow Data Export (single/dual)
No significant impact
• NetFlow v5 versus v8: little or not impact
• NetFlow Feature Acceleration:
>200 lines of ACLs and/or Policy Based-Routing (PBR)
• NetFlow versus Sampled NetFlow on the Cisco
12000 Series Internet Routers
23% versus 3% (65,000 flows, 1:100)

• Similar CPU and throughput numbers result from configuration of both NetFlow version 5 and 9
• No change in NetFlow performance after the addition of version 9
Cisco IOS Software Releases 12.0(24)S, 12.2S, and
12.3
• CPU is slightly higher immediately following initial boot up or configuration
Caused by sending Template Flowsets to Collector

Reduce CPU and memory impact on the router, collector, or network:
• Aging timers (router)
• Sampled NetFlow (router)
• Enable NetFlow Feature Acceleration (router)
• Flow Masks (only Cat6000/7600)
• Enable on specific sub-interface (upcoming router feature)
• Aggregation schemes (v8 on router or on collector)
• Filters (router or collector)
• Data Compression (collector)
• Increase collection bucket sizes (collector)
• Collector and router can be placed on the same LAN segment
(network)

• Netflow Configuration
• Checking Netflow Configuration
• Monitoring and security
– export statistics
– protocol statistics
– top flows information (top talkers)

• Defined groups of objects
1. cnfCacheInfo
• A group of objects related to cache information and configuration stored per cache configuration.
2. cnfExportInfo
• A group of objects related to Export configuration and information.
4. cnfExportStatistics
• Provides export statistics.
5. cnfProtocolStatistics
•
Provides a summary of NetFlow cache statistics per protocol and port.
6. cnfExportTemplate
• Provides Template based Version 9 flow export information and statistic.
7. cnfTopFlows
• Provides top Netflow flows.

• Collects IPv6 flow records
• Based on Netflow Version9
• Support or both ingress and egress traffic
• "Full NetFlow" i.e. non-sampled
• Data export is still IPv4
• Available in release 12.3(7)T

• Netflow is a mature Cisco IOS feature (in
Cisco IOS since 1996)
• Netflow provides input for Accounting,
Performance, Fault, Security, and Billing
Applications
• Cisco has IETF and industry leadership
• Netflow v9 eases the exporting of additional fields
• A lot of new features have been added

• sFlow® is an industry standard technology for monitoring high speed switched networks, Juniper’s devices support it.
• similar to netflow
• NetStream from Huawei Company
• SFlow Packet:













Packet header (eg MAC,IPv4,IPv6,IPX,AppleTalk,TCP,UDP, ICMP)
Sample process parameters (rate, pool etc.)
Input/output ports
Priority (802.1p and TOS)
VLAN (802.1Q)
Source/destination prefix
Next hop address
Source AS, Source Peer AS
Destination AS Path
Communities, local preference
User IDs (TACACS/RADIUS) for source/destination
URL associated with source/destination
Interface statistics (RFC 1573, RFC 2233, and RFC 2358)

• Cisco NFC
• Arbor Peakflow
• Flow tools
• Ntop
– http://ww.ntop.org
• Etc.

• Flow-tools is library and a collection of programs used to collect, send, process, and generate reports from NetFlow data.
• Can be used together on a single server or distributed to multiple servers for large deployments.
• The flow-tools library provides an API for development of custom applications for NetFlow export versions 1,5,6 and the 14 currently defined version 8 subversions.
• Version 9 is not supported now

• flow-capture - Collect, compress, store, and manage disk space for exported flows from a router.
• flow-cat - Concatenate flow files. Typically flow files will contain a small window of 5 or 15 minutes of exports. Flow-cat can be used to append files for generating reports that span longer time periods.
• flow-fanout - Replicate NetFlow datagrams to unicast or multicast destinations. Flow-fanout is used to facilitate multiple collectors attached to a single router.
• flow-report - Generate reports for NetFlow data sets. Reports include source/destination IP pairs, source/destination AS, and top talkers.
Over 50 reports are currently supported.
• flow-tag - Tag flows based on IP address or AS #. Flow-tag is used to group flows by customer network. The tags can later be used with flow-fanout or flow-report to generate customer based traffic reports.
• flow-filter - Filter flows based on any of the export fields. Flow-filter is used in-line with other programs to generate reports based on flows matching filter expressions.
• flow-import - Import data from ASCII or cflowd format.
• flow-export - Export data to ASCII or cflowd format.

• flow-send - Send data over the network using the NetFlow protocol.
• flow-receive - Receive exports using the NetFlow protocol without storing to disk like flow-capture.
• flow-gen - Generate test data.
• flow-dscan - Simple tool for detecting some types of network scanning and Denial of Service attacks.
• flow-merge - Merge flow files in chronoligical order.
• flow-xlate - Perform translations on some flow fields.
• flow-expire - Expire flows using the same policy of flowcapture.
• flow-header - Display meta information in flow file.
• flow-split - Split flow files into smaller files based on size, time, or tags.

R1(config)# ip flow-export source Loopback0
R1(config)# ip flow-export version 5 origin-as
R1(config)# ip flow-export destination 202.112.xx.xx 9800
R1(config-if)# ip route-cache flow

• Flow-tools most useful and important command
• flow-capture -w /flows/dat -m 255.255.248.0 -E5G 0/10.0.0.1/9800
– Receive flows from the exporter at 10.0.0.1 port 9800.
Maintain 5 Gigabytes of flow files in /flows/dat. Mask the source and destination IP addresses contained in the flow exports with 255.255.248.0.
• flow-capture -w /flows/dat 0/0/9800 -S5
– Receive flows from any exporter on port 9800. Do not perform any flow file space management. Store the exports in /flows/dat. Emit a stat log message every 5 minutes.

FreeBSD1# flow-print < ft-v01.2006-09-02.134114+0800 srcIP dstIP prot sPort dPort octets pkts
202.204.79.253 202.204.239.227 6 4414 1433 48 1
202.204.79.253 202.204.239.229 6 4450 1433 96 2
202.204.79.253 202.204.239.240 6 4535 1433 48 1
202.204.79.253 202.204.239.228 6 4443 1433 48 1
202.204.79.253 202.204.239.233 6 4472 1433 96 2
202.204.79.253 202.204.239.231 6 4461 1433 48 1

% flow-cat -p /flows/dat | flow-stat
IP packet size distribution:
1-32 64 96 128 160 192 224 256 288 320 352 384 416 448 480
.000 .906 .029 .004 .002 .009 .001 .001 .004 .027 .004 .000 .000 .000 .000
512 544 576 1024 1536 2048 2560 3072 3584 4096 4608
.000 .000 .001 .001 .012 .000 .000 .000 .000 .000 .000
Packets per flow distribution:
1 2 4 8 12 16 20 24 28 32 36 40 44 48 52
.812 .157 .010 .013 .006 .001 .000 .000 .000 .000 .000 .001 .000 .000 .000
60 100 200 300 400 500 600 700 800 900 >900
.000 .001 .000 .000 .000 .000 .000 .000 .000 .000 .000
Octets per flow distribution:
32 64 128 256 512 1280 2048 2816 3584 4352 5120 5888 6656 7424 8192
.000 .754 .183 .009 .012 .015 .014 .008 .004 .002 .000 .000 .000 .000 .000
8960 9728 10496 11264 12032 12800 13568 14336 15104 15872 >15872
.000 .000 .000 .000 .001 .000 .000 .000 .000 .000 .001
Flow time distribution:
10 50 100 200 500 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
.812 .000 .000 .000 .000 .000 .001 .131 .015 .001 .004 .004 .004 .002 .001
12000 14000 16000 18000 20000 22000 24000 26000 28000 30000 >30000
.000 .001 .001 .002 .001 .000 .002 .001 .001 .000 .017

– Provide a report on top source/destination
IP pairs sorted by octets
# Fields: Total
# Symbols: Disabled
# Sorting: Descending Field 4
# Name: Source/Destination IP
#
# src IPaddr dst IPaddr flows octets packets
#
202.204.192.1 10.20.0.12 1 3720 12
202.204.192.1 10.20.0.8 3 3128 11
202.204.192.1 10.20.0.9 2 3269 11
202.204.193.1 64.84.7.4 1 390 3
202.204.204.148 221.137.69.66 3 144 3
216.186.143.246 202.204.227.118 1 144 3
202.204.79.253 202.204.239.233 1 96 2

• Most materials in this PPT is from network, thanks goes to the authors
• Any Questions?