Title: RE: [redewan] PPP Multilink Ajuda2 Corrigindo
Lista de Discuss�o Rede Wan - http://www.networkdesigners.com.br

Kevison,

Note que esta RFC � "informational", portanto, nao teria exatamente a forca de um
"Standard". De qualquer forma, � interessante o conteudo, ele explica um pouco melhor o problema do resequenciamento.
[], <O-O>

> -----Original Message-----
> From: Kevison Bentes - Bol [mailto:[EMAIL PROTECTED]]
> Sent: Wednesday, January 17, 2001 5:13 PM
> To: Lista de Discusso Rede Wan
> Subject: Re: [redewan] PPP Multilink Ajuda2 Corrigindo
>
>
> Lista de Discuss�o Rede Wan - http://www.networkdesigners.com.br
>
> Oi Br�s,
>
> Acredito que essa forma de balanceamento � padronizada.
> Veja a documenta��o abaixo.
> Em http://www.faqs.org/rfcs/rfc2991.html
>
> Grato.
>
>
> Internet RFC/STD/FYI/BCP Archives
> RFC2991
> [ Index | Search | What's New | Comments | Help ]
>
> --------------------------------------------------------------
> --------------
> ----
> Network Working Group                                        
>   D. Thaler
> Request for Comments: 2991                                   
>   Microsoft
> Category: Informational                                      
>    C. Hopps
>                                                      NextHop
> Technologies
>                                                            
> November 2000
>
> Multipath Issues in Unicast and Multicast Next-Hop Selection
>
> Status of this Memo
>
>    This memo provides information for the Internet community.  It does
>    not specify an Internet standard of any kind.  Distribution of this
>    memo is unlimited.
>
> Copyright Notice
>
>    Copyright (C) The Internet Society (2000).  All Rights Reserved.
>
> Abstract
>
>    Various routing protocols, including Open Shortest Path
> First (OSPF)
>    and Intermediate System to Intermediate System (ISIS), explicitly
>    allow "Equal-Cost Multipath" (ECMP) routing.  Some router
>    implementations also allow equal-cost multipath usage with RIP and
>    other routing protocols.  The effect of multipath routing on a
>    forwarder is that the forwarder potentially has several
> next-hops for
>    any given destination and must use some method to choose
> which next-
>    hop should be used for a given data packet.
>
> 1.  Introduction
>
>    Various routing protocols, including OSPF and ISIS,
> explicitly allow
>    "Equal-Cost Multipath" routing.  Some router implementations also
>    allow equal-cost multipath usage with RIP and other routing
>    protocols.  Using equal-cost multipath means that if
> multiple equal-
>    cost routes to the same destination exist, they can be
> discovered and
>    used to provide load balancing among redundant paths.
>
>    The effect of multipath routing on a forwarder is that the
> forwarder
>    potentially has several next-hops for any given
> destination and must
>    use some method to choose which next-hop should be used for a given
>    data packet.  This memo summarizes current practices, problems, and
>    solutions.
>
> 2.  Concerns
>
>    Several router implementations allow multipath forwarding.  This is
>    sometimes done naively via round-robin, where each packet
> matching a
>    given destination route is forwarded using the subsequent next-hop,
>    in a round-robin fashion.  This does provide a form of load
>    balancing, but there are several problems with approaches such as
>    round-robin or random:
>
>    Variable Path MTU
>          Since each of the redundant paths may have a different MTU,
>          this means that the overall path MTU can change on a packet-
>          by-packet basis, negating the usefulness of path MTU
> discovery.
>
>    Variable Latencies
>          Since each of the redundant paths may have a
> different latency
>          involved, having packets take separate paths can
> cause packets
>          to always arrive out of order, increasing delivery
> latency and
>          buffering requirements.
>
>          Packet reordering causes TCP to believe that loss has taken
>          place when packets with higher sequence numbers arrive before
>          an earlier one.  When three or more packets are
> received before
>          a "late" packet, TCP enters a mode called
> "fast-retransmit" [6]
>          which consumes extra bandwidth (which could potentially cause
>          more loss, decreasing throughput) as it attempts to
>          unnecessarily retransmit the delayed packet(s).  Hence,
>          reordering can be detrimental to network performance.
>
>    Debugging
>          Common debugging utilities such as ping and
> traceroute are much
>          less reliable in the presence of multiple paths and may even
>          present completely wrong results.
>
>    In multicast routing, the problem with multiple paths is that
>    multicast routing protocols prevent loops and duplicates by
>    constructing a single tree to all receivers of the same group
>    address.  Multicast routing protocols deployed today
> (DVMRP, PIM-DM,
>    PIM-SM) [2] construct shortest-path trees rooted at either the
>    source, or another router known as a Core or Rendezvous Point.
>    Hence, the way they ensure that duplicates will not arise is that a
>    given tree must use only a single next-hop towards the root of the
>    tree.
>
> 3.  Requirements
>
>    In the remainder of this document, we will use the term "flow" to
>    represent the granularity at which the router keeps state
> (if at all)
>    for classes of traffic.  The exact definition of a flow
> may depend on
>    the actual implementation.  For example, a flow might be identified
>    solely by destination address, or it might be identified by (source
>    address, destination address, protocol id) triplet.  Hence
> "flow" is
>    not necessarily synonymous with the term "microflow" as used in RFC
>    2474 [7], which also includes port numbers.  Indeed, including
>    transport-layer information in the next-hop selection process can
>    actually be problematic.  For example, if packets are
> fragmented, the
>    transport-layer information may not be available in every packet.
>    Furthermore, having the choice of path depend on transport-layer
>    fields may negate the benefit of caching information such
> as MTU for
>    use in subsequent connections between the same endpoints.
>
>    All of the problems outlined in the previous section arise when
>    packets in the same unicast or multicast "flow" are split among
>    multiple paths.  The natural solution is therefore to ensure that
>    packets for the same flow always use the same path.
>
>    Two additional features are desirable:
>
>    Minimal disruption
>          When multipath is used, meaning that multiple routes
> contribute
>          valid next-hops, the chances are higher of routes being added
>          and deleted from consideration than when only the
> "best" route
>          is used (in which case metric changes in alternate
> routes have
>          no effect on traffic paths).  Since a higher number of routes
>          may actually be used for forwarding when multipath is in use,
>          the potential for packet reordering and packet loss due to
>          route flaps can be much greater than when not using
> multipath.
>          Hence, it is desirable to minimize the number of active flows
>          affected by the addition or deletion of another next-hop.
>
>    Fast implementation
>          The amount of additional computation required to forward a
>          packet should be small.  For example, when doing round-robin,
>          this computation might consist of incrementing (modulo the
>          number of next-hops) a next-hop index.
>
> 4.  Solutions
>
>    We now provide three possible methods for improving the performance
>    of multipath and then discuss their applicability to unicast and
>    multicast forwarding.
>
>    Modulo-N Hash
>          To select a next-hop from the list of N next-hops, the router
>          performs a modulo-N hash over the packet header fields that
>          identify a flow.  This has the advantage of being
> fast, at the
>          expense of (N-1)/N of all flows changing paths whenever a
>          next-hop is added or removed.
>
>    Hash-Threshold
>          The router first selects a key by performing a hash over the
>          packet header fields that identify the flow.  The N next-hops
>          have been assigned unique regions in the hash
> function's output
>          space.  By comparing the hash value against region boundaries
>          the router can determine which region the hash value
> belongs to
>          and thus which next-hop to use.  This method has the
> advantage
>          of only affecting flows near the region boundaries (or
>          thresholds) when next-hops are added or removed.  For ECMP
>          hash-threshold's lookup can be done with a simple division
>          (hash_value / fixed_region_size).  When a next-hop
> is added or
>          removed, between 1/4 and 1/2 of all flows change paths.  An
>          analysis of this method can be found in [3].
>
>    Highest Random Weight (HRW)
>          The router computes a key for EACH next-hop by performing a
>          hash over the packet header fields that identify the flow, as
>          well as over the address of the next-hop.  The router then
>          chooses the next-hop with the highest resulting key
> value [4].
>          This has the advantage of minimizing the number of flows
>          affected by a next-hop addition or deletion (only
> 1/N of them),
>          but is approximately N times as expensive as a modulo-N hash.
>
>    The applicability of these three alternatives depends on (at least)
>    two factors: whether the forwarder maintains per-flow
> state, and how
>    precious CPU is to a multipath forwarder.
>
>    Some routers may maintain per-flow state for reasons other than for
>    supporting multipath.  For example, routers typically keep per-flow
>    state for multicast flows so that they can maintain the list of
>    interfaces to which packets in the flow should be copied.
>
>    If per-flow state is maintained in a multipath forwarder, then
>    computation of the next-hop can be done by the router at state
>    creation time.  This entails no additional computations at packet
>    forwarding time compared with normal forwarding to a
> single next-hop,
>    since the next-hop is precomputed.  In this case, any method can be
>    used, including round-robin, random, modulo-N,
> hash-threshold or HRW.
>    Hash functions such as modulo-N, hash-threshold and HRW
> are better if
>    the forwarder state may be deleted for any reason during
> the lifetime
>    of a flow since subsequent next-hop computations by the router will
>
>    always select the same path.  This also improves the usefulness of
>    debugging utilities such as traceroute.  Finally, to maximize the
>    stability of paths (and hence the usefulness of traceroute, etc.),
>    the use of HRW is recommended over the other methods mentioned
>    herein.
>
>    If per-flow state is not maintained by the forwarder, then using
>    multiple next-hops requires that the next-hop be
> calculated at packet
>    arrival time.  When CPU is more precious than stability of flow
>    paths, hash-threshold is recommended over the other
> methods mentioned
>    herein.
>
> 4.1.  Unicast Forwarding
>
>    Depending on the implementation, unicast forwarding may or may not
>    keep per-flow state.  We recommend that where forwarder
>    implementations keep flow state, routers should use HRW at state
>    creation time (and next-hop deletion time) to select the next-hop,
>    and that forwarders without per-flow state use hash-threshold.
>
> 4.2.  Multicast Forwarding
>
>    Today's multicast forwarding engines use a cache of forwarding
>    entries indexed by group (or group prefix) and source (or source
>    prefix).  This means that today's multicast forwarder's always keep
>    per-flow state, although for some multicast routing protocols, the
>    "flow" may be fairly coarse (e.g., traffic from all sources to the
>    same destination).  Since per-flow state is kept by the
> forwarder, it
>    is recommended that the router always use HRW to select
> the next-hop.
>
>    Routers using explicit-joining protocols such as PIM-SM [5] should
>    thus use the multipath information when determining to
> which neighbor
>    a join message should be sent.  For example, when multiple
> next-hops
>    exist for a given Rendezvous Point (RP) toward which a (*,G) Join
>    should be sent, it is recommended that HRW be used to select the
>    next-hop to use for each group.
>
> 5.  Applicability
>
>    The algorithms discussed above (except round-robin) all
> rely on some
>    form of hash function.  Equal flow distribution is
> achieved when the
>    hash function is uniformly distributed.  Since the
> commonly used hash
>    functions only become uniformly distributed when the
> number of inputs
>    is relatively large, these algorithms are more applicable
> to routers
>    used to route many flows, than in, for example, a small business
>    setting.
>
> 6.  Redundant Parallel Links
>
>    A related problem occurs when multiple parallel links are used
>    between the same pair of routers.  A common solution is to
> bundle the
>    two links together into a "super"-link when is then used
> for routing.
>    For multicast forwarding, this results in the two links
> being reduced
>    to a single next-hop (over the combined link) which can be used to
>    prevent duplicates.  When a unicast or multicast packet is
> queued to
>    the combined link, some method, such as those discussed earlier, is
>    still required to determine the physical link on which to transmit
>    the packet.  If the parallel links are identical, then most of the
>    concerns discussed in this document are avoided with the combined
>    link.  The exception is packet reordering, which can still
> occur with
>    round-robin, adversely affecting TCP.
>
> 7.  Security Considerations
>
>    This document discusses issues with various methods of choosing a
>    next-hop from among multiple valid next-hops.  As such, it does not
>    directly impact the security of the Internet infrastructure or its
>    applications.
>
>    One issue that is worth mentioning, however, is that when next-hop
>    selection is predictable, an attacker can synthesize traffic that
>    will all hash the same, making it possible to launch a denial-of-
>    service attack that overloads a particular path.  Since a special
>    case of this is when the same (single) next-hop is always selected,
>    such an attack is easiest when multipath is not being used.
>    Introducing multipath routing can make such an attack more
> difficult;
>    the more unpredictable the hash is, the harder it becomes
> to conduct
>    a denial-of-service attack against any single link.
>
> 8.  References
>
>    [1]  Moy, J., "OSPF Version 2", STD 54, RFC 2328, April 1998.
>
>    [2]  Maufer, T., "Deploying IP Multicast in the Enterprise",
>         Prentice-Hall, 1998.
>
>    [3]  Hopps, C., "Analysis of an Equal-Cost Multi-Path
> Algorithm", RFC
>         2992, November 2000.
>
>    [4]  Thaler, D., and C.V. Ravishankar, "Using Name-Based
> Mappings to
>         Increase Hit Rates", IEEE/ACM Transactions on Networking,
>         February 1998.
>
>    [5]  Estrin, D., Farinacci, D., Helmy, A., Thaler, D., Deering, S.,
>         Handley, M., Jacobson, V., Liu, C., Sharma, P. and L. Wei,
>         "Protocol Independent Multicast-Sparse Mode (PIM-SM): Protocol
>         Specification", RFC 2362, June 1998.
>
>    [6]  Allman, M., Paxson, V. and W. Stevens, "TCP
> Congestion Control",
>         RFC 2581, April 1999.
>
>    [7]  Nichols, K., Blake, S., Baker, F. and D. Black.,
> "Definition of
>         the Differentiated Services Field (DS Field) in the IPv4 and
>         IPv6 Headers", RFC 2474, December 1998.
>
> 9.  Authors' Addresses
>
>    Dave Thaler
>    Microsoft
>    One Microsoft Way
>    Redmond, WA  98052
>
>    Phone: +1 425 703 8835
>    EMail: [EMAIL PROTECTED]
>
>    Christian E. Hopps
>    NextHop Technologies, Inc.
>    517 W. William Street
>    Ann Arbor, MI 48103-4943
>    U.S.A
>
>    Phone: +1 734 936 0291
>    EMail: [EMAIL PROTECTED]
>
> 10.  Full Copyright Statement
>
>    Copyright (C) The Internet Society (2000).  All Rights Reserved.
>
>    This document and translations of it may be copied and furnished to
>    others, and derivative works that comment on or otherwise
> explain it
>    or assist in its implementation may be prepared, copied, published
>    and distributed, in whole or in part, without restriction of any
>    kind, provided that the above copyright notice and this
> paragraph are
>    included on all such copies and derivative works.  However, this
>    document itself may not be modified in any way, such as by removing
>    the copyright notice or references to the Internet Society or other
>    Internet organizations, except as needed for the purpose of
>    developing Internet standards in which case the procedures for
>    copyrights defined in the Internet Standards process must be
>    followed, or as required to translate it into languages other than
>    English.
>
>    The limited permissions granted above are perpetual and will not be
>    revoked by the Internet Society or its successors or assigns.
>
>    This document and the information contained herein is
> provided on an
>    "AS IS" basis and THE INTERNET SOCIETY AND THE INTERNET ENGINEERING
>    TASK FORCE DISCLAIMS ALL WARRANTIES, EXPRESS OR IMPLIED, INCLUDING
>    BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF THE INFORMATION
>    HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED WARRANTIES OF
>    MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE.
>
> Acknowledgement
>
>    Funding for the RFC Editor function is currently provided by the
>    Internet Society.
>
>
>
> --------------------------------------------------------------
> --------------
> ----
>
> [ Index | Search | What's New | Comments | Help ]
> Comments/Questions about this archive ? Send mail to
> [EMAIL PROTECTED]
>
>
>
>
>
> ----- Original Message -----
> From: Bras, Roberto Carlos De (Bras) <[EMAIL PROTECTED]>
> To: Lista de Discuss�o Rede Wan <[EMAIL PROTECTED]>
> Sent: Friday, January 12, 2001 1:16 PM
> Subject: RE: [redewan] PPP Multilink Ajuda2 Corrigindo
>
>
> Lista de Discuss�o Rede Wan - http://www.networkdesigners.com.br
>
> Esta forma de balanceamento e padrao?
> Ou seja, existe RFC, STD ou equivalente que regule como deva
> funcionar este
> balanceamento?
>
> Ou isto depende de cada fabricante (ex. Cisco)?
>
>
> -----Original Message-----
> From: Antonio Carlos Pina [mailto:[EMAIL PROTECTED]]
> Sent: Thursday, January 11, 2001 2:04 PM
> To: Lista de Discussco Rede Wan
> Subject: Re: [redewan] PPP Multilink Ajuda2 Corrigindo
>
>
> Lista de Discuss�o Rede Wan - http://www.networkdesigners.com.br
>
> Prezado,
>
> N�o tenho certeza se entendi direito o que voc� quer. Mas se
> voc� tem 2
> links ligados na mesma rede e no mesmo roteador, para
> balance�-los basta
> criar 2 rotas com a mesma m�trica para a mesma rede.
>
> Suponha que voc� possua na outra rede o bloco 10.0.0.0/24.
> Voc� far� assim:
>
> ip route 10.0.0.0 255.255.255.0 serial 0
> ip route 10.0.0.0 255.255.255.0 serial 1
>
> E no outro roteador, ser�:
>
> ip route 0.0.0.0 0.0.0.0 serial 0
> ip route 0.0.0.0 0.0.0.0 serial 1
>
> Os roteadores balancear�o, enviando 1 pacote por uma serial e
> outro pela
> outra e assim por diante.
>
> Cordialmente,
> Antonio Carlos Pina
> Diretor de Tecnologia
> INFOLINK Internet
> http://www.infolink.com.br
>
> ----- Original Message -----
> From: ChicoZeni
> To: Lista de Discuss�o Rede Wan
> Sent: Thursday, January 11, 2001 8:40 AM
> Subject: [redewan] PPP Multilink Ajuda2 Corrigindo
>
>
> Lista de Discuss�o Rede Wan - http://www.networkdesigners.com.br
> Corrigindo...
> Ilustres companheiros da Lista, preciso configurar em roteadores CISCO
> 2500/2600/4500 dois links seriais de 256k de forma que fa�am
> balanceamento
> de carga, em PPP Multilink.
>
> router 1                                                    router2
> s1 ppp                                                   s1 ppp
> s2 ppp                                                   s2 ppp
>
> de forma que haja balanceamento de carga e que se um link cair o outro
> assuma tudo.
>  end ip 198.168.x.x1     para router 1
>  end ip 198.168x.x2     para router 2
>
> Se alguem ja tem uma copia dessa configura��o me enviar pela
> lista ou via
> private no [EMAIL PROTECTED]
>
> Obrigado aos amigos da lista.
>
> FRANCISCO ZENI
> 41 351-8589
>
>
> ______________________________________________________________________
> To unsubscribe, write to [EMAIL PROTECTED]
>


To unsubscribe, write to [EMAIL PROTECTED]

Responder a