Net Worker World

Queueing is for Congestion Management. Custom Queueing: With CQ, bandwidth is allocated proportionally for each different class of traffic, By default, each cycling, dequeque 1500 bytes for each queue. CQ allows you to specify the number of bytes or packets to be drawn from the queue http://www.ccietalk.com/2008/12/15/custom-queueing Configuration: queue-list command, then apply under interface using custom-queue-list • Defining the Custom Queue List (Required) • Specifying the Maximum Size of the Custom Queues (Optional) • Assigning Packets to Custom Queues (Required) • Monitoring Custom Queue Lists (Optional) Class-Based Weighted Fair Queueing (CBWFQ) Class-based weighted fair queueing (CBWFQ) extends the standard WFQ functionality to provide support for user-defined traffic classes. For CBWFQ, you define traffic classes based on match criteria including protocols, access control lists (ACLs), and input interfaces. First In First Out (FIFO) Queueing First In, First Out (FIFO) Queuei

Traffic Shaping doesn't drop voilated traffic, it typically delay the excess traffic using buffer or queueing mechanism. Traffic shaping controls the speed of traffic leaving an interface. Cisco has three types traffic shaping: Class-Based Traffic Shaping, Generic Traffic Shaping (GTS), and Frame Relay Traffic Shaping (FRTS). Generic Traffic Shapping: use traffic-shape command at interface level. Frame Relay Traffic Shaping: can shape traffic per DLCI. 1. Difine map-class with frame-relay command 2. use frame-relay traffic-shaping command at phsical interface to enable Frame Relay Traffic Shaping. 3. at subinterface, call class in frame-relay interface-dlci command.

Priority Queuing has four queues: High, Medium, Normal and Low. High Queue must be empties before the router serves Medium Queue, then Medium Queue must be emptied before serve Normal Queue, and so on. Configuration: Priority-list Then in interface config mode: Priority-group Verify: Show queueing References: http://www.cisco.com/en/US/docs/ios/12_1/qos/configuration/guide/qcdconmg.html#wp1001400 http://www.cisco.com/en/US/docs/ios/12_1/qos/configuration/guide/qcdpq.html

Here is the script to ping many IPs from a router, run this in tclsh. foreach IP { Ip1 Ip2 Ip3 ... IP100 }{ping $IP}

on a multiaccess LAN, DR is responsible for "PIM" Register/Prune messages toward RP. IGMP Querier is responsible for IGMP "host-query" messages. With IGMPv1, since there is no concept of Querier, PIM DR typically serves both the PIM routers and IGMP hosts (think of it as an additional responsibility). With IGMPv2, there IS an election, so router with lower ip address serves the hosts (as an IGMP Querier) and router with higher ip address serves PIM routers (as PIM-DR). The winner of Assert process is responsible for forwading multicast onto LAN

This command simulates the end client to join a multicast group, it responses to PING, the router executes PING simulates the multicast Source.

In PIM Sparse mode, a router needs send Join message to the RP to receive multicast traffic. With AutoRP, the mapping agent send RP info using 224.0.1.40, however, since PIM is in sparse mode, the 224.0.1.40 multicast won't flood in entire network, other router wants to receive this traffic but isn't aware of any RP to send join message, this is chicken-and-egg problem. There are two solutions: 1. PIM Sparse-Dense mode a router uses Dense mode if it has no info where the RP is, so 224.0.1.40 traffic can flood entire network in dense mode, then the router switch to Sparse mode since it knows RP. 2. PIM autorp listener. This command causes the router flood 224.0.1.40 traffic out all interfaces. AutoRP is enabled on the Cisco router by default, so all routers have the following mroute by default: (*, 224.0.1.40), 00:07:57/00:00:07, RP 0.0.0.0, flags: DP   Incoming interface: Null, RPF nbr 0.0.0.0   Outgoing interface list: Null In case we don't use AutoRP, we can

Tc -- Time interval, Tc=Bc/CIR Bc -- Committed burst size, the bits can be sent per Tc interval. It specifies in bits (or bytes). Also called Sustained. CIR -- Committed information rate, the bits can be sent per second. Also called Mean rate. So Bc and CIR are same thing but measured in different way. Be -- Excess burst size, bits can be sent per second beyond Bc after a period of inactivity. Shaped rate -- bits per second, it may or may not be set to CIR. At service provider side, CIR means the "guaranteed” rate of the circuit based on the SLA. At client side, “CIR” in the context of shaping simply means the target average rate. The actually speed of sending date can only be the line speed, to shape or policer the date to a lower speed, for example, to 10M of 100M Ethernet interface, the way the interface does is to send the data at line speed a very short period then stop, this send/stop is a interval called Tc, which can be 1/8 to 1/125 second. Shorter interval m

1. When a multicast application launches on a host, a IGMP Report sends out, this is join message. 2. A multicast router periodically sends out IGMP General Query to 224.0.0.1. 3. A multicast host response the query by sending out IGMP Report to the group address it is listening on. 4. On a LAN, a host uses MRT to suppress its IGMP Report when it sees other host sends IGMP report already. 5. A host sends Leave message when it is no longer interested in recevieing multicase traffic, the router will sends Group-Specific Query when receives the Leavel message, and expecting any hosts still in the group response with IGMP Report 6. If multiple routers on a LAN, the router with lowest IP Address is elected as IGMP querier.