itsmydamnations CCIE study journal

Discussion in 'Networking, Telephony & Internet' started by itsmydamnation, Sep 1, 2008.

  1. fR33z3

    fR33z3 Member

    Joined:
    Jul 16, 2001
    Messages:
    2,164
    Location:
    Perth
    time to revive an old thread....

    My motivation is starting to build up for a journey into CCIE R+S Land.

    Just the planning stage at the moment.

    Hows your study going itsmydamnation? You going for a resit soon?

    Anyone else midway through their study?
     
  2. malbert

    malbert Member

    Joined:
    Sep 19, 2009
    Messages:
    206
    Location:
    Canberra
    I'm not sure if this has been posted elsewhere, but OEQs have been dropped from R&S and Voice as of 10th May 2010. Might be time to give it another crack itsmydamnation.
     
  3. OP
    OP
    itsmydamnation

    itsmydamnation Member

    Joined:
    Apr 30, 2003
    Messages:
    10,415
    Location:
    Canberra
    yeah maybe, but right now i am full on work wise.

    cheers
     
  4. OP
    OP
    itsmydamnation

    itsmydamnation Member

    Joined:
    Apr 30, 2003
    Messages:
    10,415
    Location:
    Canberra
    so yesterday i went to the computer fair and spend 400 bux and got the following:

    Am3 MB
    amd X4 640
    4gig ddr3
    case + 500watt power supply
    20gig hdd (for $7)

    that is now my new dynamips box, i have secured the following switches:

    2X 3550 48 port
    1X 3560 G 24 port
    1X 3750 48 port
    1X 4500 with SUP V

    so after everyone from nimmers to my boss to my wife telling me i should resit the CCIE i am. I will be re-doing all of my notes as there isn't enough detail on the process of things ( like explaining how the OSPF database should look). With there now being a troubleshooting section this will be very important.

    cheers
     
  5. biatch

    biatch Member

    Joined:
    Jun 18, 2002
    Messages:
    1,679
    Location:
    North Brisbane
    Hi itsmydamnation,

    Sorry to post off-topic in your thread, but if you're studying your CCIE then it seems very fitting to ask you. :)
    I have some queries re: tcp flow control/windowing and how specifically it reacts if the sender either doesn't receive an ack or receives an ack with a sequence number lower than the ack that it sent.
    There's a brief discussion in the Billion thread which is what I'm referring to, but I'm after some clarification if possible?

    If packets are dropped (in this instance due to rate-limiting) and an ack isn't received, or is a lower seq number, how does flow control specifically respond to that? Does it send smaller chunks of data which results in lots of tiny packets, or do the packets get fragmented? Or would it still be capable of sending a full 1500bytes without fragmenting it? Presuming it's a say a single file being downloaded.
    I'm now a bit confused and doubting my understanding of it (which is no surprise to me), so would love some further explanation if you wouldn't mind?
     
  6. OP
    OP
    itsmydamnation

    itsmydamnation Member

    Joined:
    Apr 30, 2003
    Messages:
    10,415
    Location:
    Canberra
    well its been a while since i looked but it would be silly to have a window size smaller then 1500 bytes. routers/etc should never fragement packets these days thoughput is to high to waste processor time.
     
  7. evilasdeath

    evilasdeath Member

    Joined:
    Jul 24, 2004
    Messages:
    4,863
    TCP sets the initial window size to 2x MSS from memory and then increases by 1 MSS per ack received until loss occurs.

    TCP breaks up data by size MSS and never resorts to shrinking a packet down. thats up to the upper layers to decide if they want smaller packets eg VOIP etc.

    TCP sequence numbers and ordering as far as i know are irreverent they can arrive out of order and it won't affect flow control. Loss however will result in retransmissions and also a change in the window size.

    How the window size changes depends very much on the version of TCP being used there are several, and you can customize it yourself if your game. But everything is controlled by the host so if your using linux/BSD etc you can compile your own TCP stack if you wish.
     
  8. biatch

    biatch Member

    Joined:
    Jun 18, 2002
    Messages:
    1,679
    Location:
    North Brisbane
    Yes I thought the same, but now I'm doubting my understanding of it.

    The other person's thought was that tcp's sliding window would create smaller than 1500byte packets due to dropped packets/acks not received, or fragmenting the packet due to a 64kbit rate-limit. At least that's my interpretation of what they were trying to say.

    If you please have a read over it I'd be appreciative. I may be way off base, in which case I'd better go over the mechanism again!
     
    Last edited: Sep 19, 2010
  9. OP
    OP
    itsmydamnation

    itsmydamnation Member

    Joined:
    Apr 30, 2003
    Messages:
    10,415
    Location:
    Canberra
    i really dont think it would, a 64k rate limit is quite bad, depending what they mean by rate limit ( traffic shaping/ traffic policing, or "rate-limit" ) the effects on a 1500 byte packet will change. how many internet facing routers allow fragmented packets? fragmenting packets would be a silly thing to do.

    i think its stupid that tcp windowing uses bytes it should use number of packets, this would have also allowed for much larger windows then what we currently get under normal operations.

    this si what i have in my notes for windowing:

     
    Last edited: Sep 19, 2010
  10. Heywood

    Heywood Member

    Joined:
    Dec 25, 2001
    Messages:
    457
    Your lab set up is similar to mine (though I have a 630 in my dynamips box) and have 3 x 3550s and 2 x 3750s - I'm sure you'll find it does the job quite well.

    Good to know you're back in the game - getting the push and support from home is great. Is your boss subsidising anything/letting you have some down time to study?

    The troubleshooting part will be interesting to face - something like between 8 and 12 tickets to solve in 2 hours on a virtualised network larger than the config portion of the lab.
     
  11. nimmers

    nimmers Member

    Joined:
    Dec 20, 2005
    Messages:
    1,193
    Location:
    Sydney
    Mate you failed on OEQ and passed the prac, I reckon you should kill it now that OEQ are gone.
     
  12. OP
    OP
    itsmydamnation

    itsmydamnation Member

    Joined:
    Apr 30, 2003
    Messages:
    10,415
    Location:
    Canberra
    updated my OSPF document added some information about how the ospf databse looks. all i can say is that type 1 and type 2 router LSA's are stupid type 1 has the link details bar the prefix length and type two has the DR RID and prefix length. I even did a packet capture and thats the way it is in the LSU's.

    So that means to actually look up a prefix in the ospf database you need to look up both the type 1 and 2 but neither of them clearly show what the prefix is.


    should note that this is fixed in OSPF V3 as there is now a "link LSA" (type 8)
     
    Last edited: Sep 21, 2010
  13. evilasdeath

    evilasdeath Member

    Joined:
    Jul 24, 2004
    Messages:
    4,863
    LSA not SLA, might want to work on your dyslexia :)
     
  14. OP
    OP
    itsmydamnation

    itsmydamnation Member

    Joined:
    Apr 30, 2003
    Messages:
    10,415
    Location:
    Canberra
    i got LSU right :lol:

    right nw my plan is to nock off the "route" exam and QOS exam so i can get CCIP, then its CCIE written again and then lab. So first im going over routing theory, then QOS, even everythig else for written. from there i will book a lab and just do losts of prac labs (IEWB VOL1/2/4 will be my tools). the plan is to get the written ASAP so i can lock in my lab date.

    cheers
     
    Last edited: Sep 21, 2010
  15. evilasdeath

    evilasdeath Member

    Joined:
    Jul 24, 2004
    Messages:
    4,863
    CCIE question 1

    Configure IP SLA

    itsmydamnations answer

    configures OSPF :)

    dyslexia well get you, don't worry it mess's with me all day reading acronyms and various numbers.
     
  16. OP
    OP
    itsmydamnation

    itsmydamnation Member

    Joined:
    Apr 30, 2003
    Messages:
    10,415
    Location:
    Canberra
    im only dsylexic in writting, when i was typing that out i was thinking LSA.
     
  17. OP
    OP
    itsmydamnation

    itsmydamnation Member

    Joined:
    Apr 30, 2003
    Messages:
    10,415
    Location:
    Canberra
    stupid eigrp

    stupid eigrp!!!

    have a look at this for fun!!!

    and now the topology table


    but then

    So the links are balanced but the second link doesn't meet the FC but because metric is equal both paths enter the table yet your bog standard show ip eigrp topology doesn't show the second path.

    YAY!
     
  18. MrvNDMrtN

    MrvNDMrtN Member

    Joined:
    Dec 24, 2001
    Messages:
    1,355
    Location:
    SW Syd
    maybe the topology table doesnt show more than 1 entry?

    The FIB and detail should be the only things that matter?
     
  19. OP
    OP
    itsmydamnation

    itsmydamnation Member

    Joined:
    Apr 30, 2003
    Messages:
    10,415
    Location:
    Canberra
    if there are FS's then they are listed in the topology table under the F route. its just one of those odd situations because normaly for eigrp multipath you wont have paths with equal cost but very close cost and with the varency command entered you will have slightly unequal cost load balancing.

    the topology table does matter because its a basis for part of the RIB, the RIB doesn't determine the best path or a loop free topology for eigrp the topology + "DUAL" does that, having the routing table not match the topology table output for routes that are in the RIB is a pain
     
    Last edited: Sep 28, 2010
  20. evilasdeath

    evilasdeath Member

    Joined:
    Jul 24, 2004
    Messages:
    4,863
    So its just a display issue then, because both routes have landed in your routing table.

    10.1.4.0/30 has done the same thing also.

    Its pretty clear thou when you read the line that says 2 successors that more information is available its just not shown.

    I do find interesting after a quick google

    http://www.cisco.com/en/US/docs/ios/12_3/iproute/command/reference/ip2_s2g.html#wp1042641

    This output shows a slightly different layout and result from sh ip egirp top, so it may be different through different IOS revisions.

     

Share This Page

Advertisement: