Network Security Wiki

BGP OSPF Questions: Difference between revisions

Page Discussion

BGP OSPF Questions (view source)

Revision as of 21:11, 4 November 2022

12,559 bytes removed ,  1 year ago
m
→‎load balacing mehtods
 
(4 intermediate revisions by the same user not shown)
Line 20:
*Medium branch - up to 100 users .for medium/large we should have mutlilayer architecture to provide high availiblity and resilency,
*Large branch - up to 200 users or more
 
= Queing - FIFO,PQ,WFO,CBWFQ =
 
* PQ - High priorty queue is always serviced first irrrespective traffic coming fron other queue.
* WFQ - Flow based ,each flow consist of source port ,destination port ,source and destination WFO always give prefernce smaller flows and lower packet size
* CBWFQ - Each traffic is classifed and placed in class ,each class is allcated some amount of bandwidth ,queues are always serviced on basis amount of allocated bandwidth to queue .
 
* Random Early Detection (RED) is a congestion avoidance mechanism that takes advantage of the congestion control mechanism of TCP. By randomly dropping packets prior to periods of high congestion, RED tells the packet source to decrease its transmission rate. WRED drops packets selectively based on IP precedence. Edge routers assign IP precedences to packets as they enter the network. (WRED is useful on any output interface where you expect to have congestion. However, WRED is usually used in the core routers of a network, rather than at the edge.) WRED uses these precedences to determine how it treats different types of traffic.
 
* When a packet arrives, the following events occur:
1. The average queue size is calculated.
2. If the average is less than the minimum queue threshold, the arriving packet is queued.
3. If the average is between the minimum queue threshold for that type of traffic and the maximum threshold for the interface, the packet is either dropped or queued, depending on the packet drop probability for that type of traffic.
4. If the average queue size is greater than the maximum threshold, the packet is dropped.
 
= IPSEC =
Line 865 ⟶ 851:
switch(config-vsan-db)# vsan <number> interface vfc <number>
switch(config-vsan-db)# exit
 
= F5 Trainging =
 
LTM How BIG IP process Traffic
 
 
Node -represent the Ip address
Pool member -combination of Ip address and port number ,in other words pool member is application server on which F5 will redirect the traffic
Pool-combitnation of pool memeber.
 
Virtual server -combination of virtual IP and port ,is also know as listner and we associate virtual server to pool members.
 
= load balacing mehtods =
static -Round robin ,ratio
Dyanamic -LFOPD (least connection ,fastest ,observed,predective,dyanmic ratio )
 
 
 
Least connection -load balacing is based on no of connection counts ,if the connection counts are equal it will use round robin
 
 
Fastest -No of layer 7 request pending on each member.
 
Observed -ration load balacing method but ratio assigned by BIG IP,No off least connections counts BIG IP assign the request and check dyanamically and assign the ratio's of the request.
 
Predective -similar to oberved but assigns the ratio agressivley based on average connection counts .
 
 
load balacing by poolmember or node .
 
 
Priorty activation -helps to configure back sets for exsiting pool members .BIG Ip will use high priorty pool member first .
 
Fallback host is only used for HTTP request ,if all the pool memebers are not availiable BIG will redirect the cilent request
 
--------------------------------------------------------------------------------------
 
Monitors :check the status of nodes and pool memembers ,if any pool meember resposnse time is not good or is not responding big ip will not send the request to that node.
 
monitor type :
 
adress check -BIG IP send ICMP request and wait for reply if there is no reply it considers nei down does not send the trafic further to that node.
 
service check -will check TCP port number on which server is lisenting ,if no responce it considers down ----
 
contect check -we can check if the server is resondping with right contest ,like for http requset get/http .... request is send .
 
interactive check -TEST for FTP connection .once connection is open username and password is send then request is send get /file once file is recieved connection is closed .
 
F5 recommends time out = 3n+1 (frquency) for setting the monitor for http
 
Customization of monitor
 
Assign nodes to monitor
 
 
-------------------------------
 
Profiles -defining traffic behaviour for virtual server.
 
Profiles contains setting how to process traffic though virtual servers.if for certain application BIG IP load balace the traffic then it will break the client connection
to avoid this we use perstiance profile so that return request for the cilent is send to same server.
 
persisteance profile - isconfigured for clients and group of cilents how BIG IP knows the returning client request need to send to same server ,persistance profile is confiured taking source ip address of http cookie
 
SSL termination -
 
 
FTP profile
 
 
All virtual servers have layer four profile includes tCP,UDP,fastl4
 
 
Profile types -service profile ,persistance profile ,protocol profile ,ssl profile ,authentication profile ,other profiles.
 
 
persistence types-
----------------------------------------------------------------
 
source address persistance :keeps the track of source ip address ,adminstrator can set the net mask in persitance record so that all lients in same mask will assigned to same pool member.
 
Limitation -if the client address being NAted .
 
 
Cookie persistance -only uses http protocol
 
Three modes : (insert ,rewrite ,passive ) mode.
 
Insert mode -BIG ip create special cookie in HTTP resonse to client .
rewrite -pool member created blanl cookie and big ip inserts special cookie
passive -pool memeber created special cookie and BIG IP let it pass through
 
-------------------------------------------------------------------------------
 
 
SSL Profile
 
SSL is secured socket layer .
 
website which uses HTTPS we need to us SSL profile as traffic is being Nated for source clients and web app is using https protocol.
Using SSL termination BIG can decrypt the traffic and assigned to pool member.
 
 
BIG IP contains SSL encryption hardware so all the encruption and key exchange are done in hardware .centralized certifiacte management.
 
 
 
----------------------------------------------------------------------------------------
I rule :
 
is a script that direct traffic though BIG IP , based on TCl command language .I rule give controll of inbound and outbound traffic from BIg IP.
 
I rule contains follwing events ( I rule name ,events ,condtion ,action )
 
 
= Multicasting =
 
Ranges
 
224.0.0.0/4 - 224.0.0.0 -239.255.255.255
 
Link local address - 224.0.0.0/24
 
Source specifc multicast -232.0.0.0/24
 
Administrativley scoped -239.0.0.0/8
 
 
Multicast control plane work differntly than unicast routing ,it needs to know who is sender of mutlicast and to which group ,also the reciever of multicast.
 
Multicast Data plane -do RPF check ( was traffic received on correct interface and bulid multicast routing table ).
 
Multicast is source based routing .
 
IGMP -Host on LAN singanl the router to join the mutlicast group .
 
Two kind of request - (*,G)-Any source who is genrating the mutlicast stream for that group -Supported by IGMP V1 and V2
(S,G)-want to join particular source sending the mutlicast group .-IGMP version 3 support both (s,g and (*,G)
 
IGMP get enabled when the IP PIM [ Dense mode,sparse mode,SParse-DENSE-mode) is enabled .
 
BY default IGMP version 2 is enabled .
 
IP IGMP join group address can be used for testing on routers to see weather muticast traffic is recieved on router for particular group.
 
ip igmp static group command can be used to mannually put the request for particular mutlicast group insteaed of reling on IGMP queriy messsage for particular group.
 
PIM- used to siganl routers to bulid muticast tree ,tree could be sender to receiver or sender to rendevpoint--- receiver.
 
PIM version 1 or 2 ,By default its PIM version 2 , RP information is already encoded in PIM packet in version 2. PIM version 2 has field for BSR.
 
DENSE mode - Implicit join ,mutilcast traffic is send across entire network unless if some one report for not joing the particular stream.Flood and prune behiviour.
Nighbor discovery on multiicast address 224.0.0.13 same for sparse mode as well .
 
Note if we have (*,G) entery then we know about reciver and if we have (S,G) entry then we know about sender as well .
 
Two ways to genrate mutlicast traffic either through pinging mutlicast address or through IP SLA.
IN PIM dense -through RPF nei information is used to send unicast packet back to source ,message could pim prune or graft message .when the multicast source flood the traffic for particular multicast groups each multicast enable router will install (S,G entry) and (*,G) entries even if they are not intersted .
 
So in dense every router needle to install (*,G ) and (S,G) entry as we canot have (S,G) untill we have (*,G) entries.so if the source is active every router need to maintain the state table for mutlicasting .
 
Graft message for (S,G) entry is to unprune the mutlicast traffic as earlier it was set to prune .
 
State refresh to keep the link prune as its original state .
 
SParse mode -uses explict join unless it is asked by someone to join mutlicast traffic uses RP as reference point.In case we are using source specific mutlicast we don't need RP.for Group specfic joins we need RP.Traffic is not send anywhere unless it is requested .Sparse mode uses both source based trees and shortest path trees
RP needs to know the recievers and senders . DR on lan segment send (S,G) register mess age to and RP in turns reply regiester stop process and recievers on lan sengment send IGMP join and which will be converted to pim join(*,G) message to RP to form RPT tree.So pim join will traverse from receiver till RP every device will have (*,G) entry and from source till RP every device will have (S,G) entry.once RP knows about sender and reciver it will send (S,G) join request back to source and source would start sending the mutlicast traffic to RP then to receiver.then its up to the last hop reouter on reciever side for the optimation process weather it want to join directly to source using SPT bypassing RP.
 
Note -When we do debug only process switchd traffic is debug if we want to debug the data plane traffic then we need to disable cef (no ip route cache),if we change the unicast routing it will also change the mutlicasting routing,To change the unicast routing we can also use Ip mroute command .
 
 
Source based tree- tree is bulid based on shortest path from reciver till sender.
shared tree -tree from sender to RP and then RP till receiver.
 
To check RP configured on each transient router -sh ip pim rp mapping
RP can be assigned staticaly (ip pim rp address ) or dynamically ( auto RP and BSR)
 
Auto RP -uses two data plane mutlicast address (224.0.1.39) advertised by routers willing to become RP to mapping agents ,
224.0.1.40- chooses the RP and advertised to rest of routers for RP information.
 
To stay on shared tree rather than SPT ( ip pm spt-threshold infinity)
 
 
SParse-dense-mode -ANY group for which we have RP assigned used sparse mode for other uses dense mode.
 
RPF check is used for loop free path in mutlicast data plane ,AS per RPF check if the mutlicast packet is received on incoming interface router will check the unicast routing for source and that matches the incoming interface RPF check Passes else fail .
 
Once the mutlicast routing table is populated router always prefer (S,G) over (*,G) and in muticast routing table we have incoming interfaces and OIL for outgoing intefrcae list if the RPF check passes mutilcast traffic is send across all interfaces in OIL.
 
 
 
 
 
 
 
 
 
On multicast router -sh ip igmp group -- shows which multicast group is active on ethernet and which receiver has joined the group
 
To determine which router is IGMP querier router - sh ip igmp interface EO
 
We can manauly tune the query interval and query max response time -
query interval - ip igmp query interval 120 (default 60 sec)
respose time - ip igmp query-max-response-time 20 (default 10 sec)
 
IOS command to support which version of IGMP is - Ip igmp version 1/2
 
 
Test commands for IGMP
 
ip igmp join group
 
ip igmp static group
 
for sparse mode we need to assgn RP - ip pim rp address x.x.x.x
 
inorder to check if there are any rp mapping - sh ip pim rp mappings
 
Inoder to check for mutlicating packet conuters- sh ip mroute counters
 
In sparse mode there is SPT switch over shorted path tree
 
for the SPT threshold we can set the threshold on DR muticast router that is receiving the IGMP join request in gloabl config mode ip pim spt threshold (vlaue)- Value is volume of multicast feed
 
if the Rpf check is failing we can still have interface to forward multicase by static mrouter ( ip mroute server mask next hop address )
 
 
= Security =
Retrieved from "https://aman.awiki.org/wiki/Special:MobileDiff/4518...4684"