1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.

disturbance has returned

Discussion in 'Networks' started by kobem, Jan 9, 2008.

  1. kobem

    kobem Megabyte Poster

    791
    1
    50
    i know that you missed me so much , as you can see i couldn't break your heart
    and returned :p

    i am about to reach to death cos details confused my mind a little too much. I figured out
    payload length , next headers and some other facts . However , this time QoS diversity
    among ipv4 and ipv6 and stateless thing mix my mind.

    1- main issue about QoS is that in ipv4 it was provided by ToS then it was changed
    and some of you say that QoS mechanism didn't exist in ipv4 :eek:

    second one is that "Traffic class in ipv6" is said to be the referred field to ToS of ipv4
    other is flow label.As far as i can see both of them provide "prioritizing traffic"
    so what is the difference?


    2- As i said before here , "stateless indicates being "state-less" , without any help for configuration
    and it is possible if no DHCPv6 server is available in environment.

    And this type of configuration is announced as a "new feature" comes in ipv6 but it is same with APIPA.......
    so "what does it this a new feature?"


    3- While configuring a stateless ip address , can we mention all types ?(link-local(the most told
    with stateless config.), site-local
    even global..)
     
    Certifications: CCNA
  2. kobem

    kobem Megabyte Poster

    791
    1
    50
    men where are you ? i need you too much
     
    Certifications: CCNA
  3. Bluerinse
    Honorary Member

    Bluerinse Exabyte Poster

    8,871
    167
    256
    Welcome back Kobem 8)

    I can't help with your IP v6 questions as i haven't studied it yet.

    Once you have worked it all out, please be sure to post on here explaining to the vast majority of people here that dont have a clue, how it all works :)
     
    Certifications: C&G Electronics - MCSA (W2K) MCSE (W2K)
  4. kobem

    kobem Megabyte Poster

    791
    1
    50
    i am grateful to you thinking of me
     
    Certifications: CCNA
  5. Tinus1959

    Tinus1959 Gigabyte Poster

    1,539
    42
    106
    Just a quick look at http://www.rfc-editor.org/cgi-bin/rfcsearch.pl and a search on ipv6 gave me 127 RFCs with in them all answers to your questions. I did not study ipv6, but for almost any question on the technical build-up of protocols, the RFC-list must be your first source.

    Good luck.
     
    Certifications: See my signature
    WIP: MCSD, MCAD, CCNA, CCNP
  6. kobem

    kobem Megabyte Poster

    791
    1
    50
    you still suppose i do not search ... you are mistaken about that
    i have found lots of documents and can't find the answers for the questions i ask to you...
    sorry but that link doesn't help me!
     
    Certifications: CCNA
  7. greenbrucelee
    Highly Decorated Member Award

    greenbrucelee Zettabyte Poster

    14,283
    254
    329
    You can find anything on google Kobem, if you type something along the lines of what you are looking for you will find it.

    If you can find how to make a homemade grenade out of household chemicals on google, you can find the answer to your questions.

    I do not know about the subject or subjects you are studying so I can't help you.
     
    Certifications: A+, N+, MCDST, Security+, 70-270
    WIP: 70-620 or 70-680?
  8. Tinus1959

    Tinus1959 Gigabyte Poster

    1,539
    42
    106
    That is strange. RFC (request for command) are so to speak the bleueprints for the internet protocols. All fields are described in the RFCs. If a vendor would like to implement ipv6 in his OS, he would go for the RFCs to find out exactly what bit is doing what in the fields.
    Sorry Kobem, but you are mistaken.
     
    Certifications: See my signature
    WIP: MCSD, MCAD, CCNA, CCNP
  9. Cockles

    Cockles Megabyte Poster

    664
    24
    74
    What the Dickens have you been looking up dude :D
     
    Certifications: None
    WIP: Trying to find my car keys
  10. greenbrucelee
    Highly Decorated Member Award

    greenbrucelee Zettabyte Poster

    14,283
    254
    329
    I was thinking of doing a one man crusade and stop the war in Afganistan Rambo style :lol: nah twas just and example.
     
    Certifications: A+, N+, MCDST, Security+, 70-270
    WIP: 70-620 or 70-680?
  11. hbroomhall

    hbroomhall Petabyte Poster Gold Member

    6,623
    115
    224
    I'm not sure what the question is here.
    Read the RFCs (2460 and 2474 I think) for more on this. Don't forget that they are experimental. But the flow label is intended to *label* a particular stream, whereas traffic class might be used to prioritize packets within an organization.

    Who says it is a 'new feature'? It seems to me to be an advance on what was available in IPv4. And don't forget that APIPA was a very late afterthought - unilateraly produced by Microsoft, and only later did the IETF get round to producing a RFC.

    Er - what do you mean by 'mention'?

    Harry.
     
    Certifications: ECDL A+ Network+ i-Net+
    WIP: Server+
  12. Rover977

    Rover977 Byte Poster

    242
    11
    40
    Certifications: A+, Network+, Cisco CCNA
    WIP: Maths
  13. kobem

    kobem Megabyte Poster

    791
    1
    50
    1- hey harry i asked this :

    ------according to documents ipv4 was using QoS and some of you say the opposite
    which one is right?-------

    2- you asked me what i meant by "mentioning"
    and i wanted to learn stateless autoconfig. includes what type of addresses .. after a while
    from that if there is a router you can assign link-local , site-local and global. Therefore
    if not , only link-local ........

    3-due to some documents i have read , flow label field describes the "specific class of field "

    ---------note : class of field is synonym to traffic class field and additional support for real-time datagram delivery additionally QoS features ------------

    i got such a result from that >>> traffic class field is used to handling all kinds of packets though
    flow label is just for the ones which requires special handling

    --------am i thinking wrong?-------------
     
    Certifications: CCNA
  14. Spice_Weasel

    Spice_Weasel Kilobyte Poster

    254
    45
    45
    Kobem,

    Regarding QoS - IPv4 and IPv6 use exactly_the_same dscp field (just the name is changed). IPv4 has QoS support and it is widely used. Since the exact same DSCP field is used in IPv6, you can see that IPv6 also has QoS support, which will also be widely used. IPv6 also has a 20 bit flow label field - this identifies traffic flows and will be useful to some QoS configurations.

    Also, keep in mind that the header fields available are only part of QoS. Many other factors are involved, some of which have been improved upon in IPv6. So it is fair to say that some of the other improvements in IPv6 (e.g. fragmentation and encryption) can indirectly improve QoS. But strictly speaking, based on the ToS/traffic class headers, IPv4 and IPv6 are virtually identical.

    Spice_Weasel
     
    Certifications: CCNA, CCNP, CCIP, JNCIA-ER, JNCIS-ER,MCP
    WIP: CCIE
  15. hbroomhall

    hbroomhall Petabyte Poster Gold Member

    6,623
    115
    224
    I thought I'd talk to some of our networking people before I made further comments on this thread, in particular on the DSCP/TOS discussion.

    The results of this can be summarised as:
    1) TOS was never widely used. In the end it was replaced by DSCP.
    2) DSCP *is* used - but only within organizations. Very few ISPs trust the values in packets handed over to them, so will either zero the field, or ignore it
    3) As a side effect of 2) - end-users who buy routers that claim 'QOS' and set this field will not get what they expect - their ISP will usualy ignore the setting.
    4) There is no expectation that IPv6 will be any different. Basicaly most organizations will set their own policies, and will disregard such fields. The only exception to this is where a customer has included such useage in the contract.

    Harry.
     
    Certifications: ECDL A+ Network+ i-Net+
    WIP: Server+
  16. Spice_Weasel

    Spice_Weasel Kilobyte Poster

    254
    45
    45
    Harry is right, isp's do not trust dscp and typically rewrite or ignore it - if you want better service you'll have to pay for it and get an SLA covering the exactly how the isp is to treat your traffic. QoS is relevant within an organization; it is standard practice to rewrite dscp values at the network edge (e.g. the trust/untrust boundary), including internal hosts, although most company networks I've seen don't bother to rewrite dscp values unless they will be using them internally. But even when dealing with traffic destined for your isp, QoS can still be useful. For example, we had a client recently set up two mpls links to a remote site. Normal traffic can result in congestion at the mpls bottleneck; QoS is needed to prioritize, reorder, delay, drop and rate limit traffic in defined classes to ensure best use is made of the links. In another example, we maintain site to site vpn's to various clients. Congestion can be a problem, so I typically sort traffic and provide certain bandwidth allocations to ensure important traffic functions properly. The isp only provides best effort service, but QoS enables me to make best use of that service.

    QoS is more than setting bits in a field - as Harry points out, the DSCP will usually be overwritten at the next trust/untrust boundary. From my perspective the key aspects are packet reordering, shaping and policing; fragmenting and interleaving if needed (e.g. voip); allocation of bandwidth and drop priority. Sometimes I use dscp values, but mostly I use other methods of distinguishing traffic, e.g. address, port, protocol, nbar, etc. I don't use QoS everywhere (best effort is often good enough) but when it is needed it is very useful.

    Spice_Weasel
     
    Certifications: CCNA, CCNP, CCIP, JNCIA-ER, JNCIS-ER,MCP
    WIP: CCIE
  17. kobem

    kobem Megabyte Poster

    791
    1
    50
    came back!

    for my previous questions

    1- so there was no QoS but , devices sold in here tell that it has the QoS features , is it right?

    2- yeah i think class of field contains all type of packets , however , flow label is used to just
    coping with special one such as real-time types ?

    3- a new one : who produces the prefixes , a router ? (i have seen nothing related to this)
    if router produces it and "if you can only use link-local if there is no router " , so how can that prefix
    be generated?

    4-
    and "do you think the note below is right?"

    note : link - locals and site-locals are not routable on the internet , link-locals are not used to
    forward or receive data(exchange) nevertheless , site-locals used for it.
    global ones go all the way .


    router there - -- - - link-local , site-local , global
    router not exists ---- only link-local
     
    Certifications: CCNA
  18. hbroomhall

    hbroomhall Petabyte Poster Gold Member

    6,623
    115
    224
    If you mean - there are routers sold claiming to do QOS, then see my point 3 above. They may not do what you expect.
    Flow label I suspect will be used in a manner similar to MPLS. i.e. as a method of labeling a particular set of traffic, and I doubt it will be confined to 'real-time' traffic.
    Either using IPv6 ND or IPv6 DHCP
    Correct
    No idea what that means.
    Correct, assuming no DHCP exists.

    Harry.
     
    Certifications: ECDL A+ Network+ i-Net+
    WIP: Server+
  19. kobem

    kobem Megabyte Poster

    791
    1
    50
    hi again harry!

    "3- a new one : who produces the prefixes , a router ? (i have seen nothing related to this)
    if router produces it and "if you can only use link-local if there is no router " , so how can that prefix
    be generated?

    ”
    Either using IPv6 ND or IPv6 DHCP "

    ....................

    who produces global prefix site local prefix and link local prefix ?



    ...........................................................



    Quote:
    “

    Originally Posted by kobem View Post
    , link-locals are not used to
    forward or receive data(exchange) nevertheless , site-locals used for it.
    global ones go all the way .

    ”
    No idea what that means.

    no idea ? link-locals are used for just "neighbor discovery" i mean
     
    Certifications: CCNA
  20. hbroomhall

    hbroomhall Petabyte Poster Gold Member

    6,623
    115
    224
    What was wrong with my answer? Obviously a link-local address will be set up without those - the machine itself will determine it.

    Note - I see that in more recent RFCs that 'site-local' has been deprecated.

    Harry.
     
    Certifications: ECDL A+ Network+ i-Net+
    WIP: Server+

Share This Page

Loading...