Sorry, I'm currently out of the office.
If this is concerning a UCA Information Technology matter that requires
immediate attention, please contact the IT HelpDesk at 450-3107 or
helpd...@uca.edu.
If your message was personal or of a "non-urgent" nature, I'll reply as soon as
I can.
Wayne
>>> "Jolie, David" <djo...@enterasys.com> 11/25/13 07:51 >>>
As noted in a prior e-mail, Link Aggregation path selection can sometimes
favor one path so it is important to understand traffic flow in your
network when configuring the distribution algorithm. By default on
S-series the path selection is based on a hash of the Destination IP -
Source IP (DIP-SIP) pairings. This normally results in a pretty fair
distribution of traffic but, in the case of a single server sending a ton
of traffic to a single destination (i.e. the backup example mentioned
earlier), all that traffic would take a single path since it's a single
DIP/SIP pair.
For S-series you can choose an alternate distribution method as follows:
set lacp outportAlgorithm {dip-sip | da-sa | round-robin}
Also note that the distribution method effects only traffic leaving the
switch in question - you need to configure both ends of the LAG or you may
still see unequal traffic distribution.
By contrast, with MSTP you only get a single path (one per VLAN if
configured that way) and you don't have the option for round-robin with
MSTP...
Cheers,
Dave J
David Jolie
ECE-Networking, ECE-Network Security
Solutions Engineer
Extreme Networks
Cell: +1.978.436.0726
On Mon, Nov 25, 2013 at 4:50 AM, Gradelain Ngouni <
gradelain.ngo...@scaltel.de> wrote:
> Hi James,
>
>
>
> as Benjamin just stated.
>
> Ill prefer Rapid Spanning Tree and configured additionally LACP. In a
> case you have 2xS4 at the core. Just configure VSB and LACP the edge
> Switches to the core è Loop-free.
>
> But, still keep the Spanning-tree active.
>
>
>
> Regards
>
>
>
> Gradelain
>
>
>
> *Von:* Benjamin Maze (Europe) [mailto:benjamin.m...@dimensiondata.com]
> *Gesendet:* Montag, 25. November 2013 10:40
> *An:* Enterasys Customer Mailing List
> *Betreff:* RE: [enterasys] LAG vs MSTP for redundant switch links
>
>
>
> Hi James,
>
>
>
> The best choice is the Lag without any hesitation
>
> LACP will ensure you a convergence time under 300ms which is quite
> transparent for your application
>
> I thing that MSTP convergence take around 1s following the case of failure
>
>
>
> The other benefit of LACP versus Spanning tree is the stability of the
> network. The spanning tree is often a source of big problem in case of loop
> in your network.
>
>
>
> The spanning tree should be configured in addition of LACP (dont disable
> spanning tree on LACP link), that ensure a protection in case of an
> incident in LACP (like a bug)
>
>
>
> Today, the best practice in network design is to avoid spanning tree
>
>
>
> So, I recommend you to use LACP
>
>
>
> Regards,
>
>
>
>
>
>
> * Benjamin Maze*Pre-Sales Engineer
> Dimension Data France
> Tel: +33 1 4975 8626
> Mob: +33 6 6025 6678
> Fax: +33 1 4975 8629
> benjamin.m...@dimensiondata.com
>
> Zone Orlytech - 20 avenue Louis Blériot
> (adresse postale 91781 Wissous Cedex), Paray Vieille Poste, 91550, France.
>
> Dimension Data is an NTT Group company. For further information about
> Dimension Data, please go to www.dimensiondata.com
> Follow us on Social Media Blog <http://blog.dimensiondata.com>
> Facebook<http://www.facebook.com/pages/Dimension-Data-Europe/208242309194429>
> LinkedIn <http://www.linkedin.com/company/dimension-data>
> Twitter<http://twitter.com/DimensionData>
>
> P please consider the environment before printing this email
>
>
> [image: Dimension Data launches Cloud
> Services]<http://eucloud.dimensiondata.com/>
>
> *From:* James Andrewartha
> [mailto:jandrewar...@ccgs.wa.edu.au<jandrewar...@ccgs.wa.edu.au>]
>
> *Sent:* lundi 25 novembre 2013 02:58
> *To:* Enterasys Customer Mailing List
> *Subject:* [enterasys] LAG vs MSTP for redundant switch links
>
>
>
>
>
> Hi list,
>
> I'm setting up dual fibre links between our core and edge switches, and
> was pondering whether to set up link aggregation or use MSTP to balance
> the traffic across the links. My primary concern is for redundancy, not
> extra bandwidth (I just checked cacti, and most don't sustain 100Mb/s).
> How do MSTP and LACP compare for failover times?
>
> The other thing about LACP is the config overhead of having to set the
> aadminkey on the physical and LAG ports plus ensuring the VLANs match.
> Whereas with MSTP I don't have to worry about the LAG port, and can just
> set a port priority on the SID to balance the traffic if required.
>
> I'm leaning towards MSTP, but every man and his dog seems to have a
> spanning-tree meltdown story. All the switches will be Enterasys (S4 at
> the core, B3/B5 at the edge) so you'd think it should all work fine.
> Opinions?
>
> Thanks,
>
> --
> James Andrewartha
> Network & Projects Engineer
> Christ Church Grammar School
> Claremont, Western Australia
> Ph. (08) 9442 1757
> Mob. 0424 160 877
>
> ---
> To unsubscribe from enterasys, send email to lists...@unc.edu with the
> body: unsubscribe enterasys benjamin.m...@dimensiondata.com
>
>
> itevomcid
>
> - --To unsubscribe from enterasys, send email to lists...@unc.edu with
> the body: unsubscribe enterasys gradelain.ngo...@scaltel.de
>
>
>
> *Gradelain Ngouni*
> Dipl. -Ing.
> IT-Projektleiter
>
>
> SCALTEL AG
> Anna-Birle-Str. 2
> 55252 Mainz-Kastel
>
>
> Telefon +49 6134 50789-23 Telefax +49 (0) 6134 50789-10
> gradelain.ngo...@scaltel.de
> Rechtsform: Aktiengesellschaft
> Registergericht: Kempten HRB 7208
> Sitz: Waltenhofen
> Vorstandsvorsitzender: Christian Skala
> Vorstand: Joachim Skala
> Aufsichtsratsvorsitzender: Alfons Hörmann
>
> *SCALTEL-Film:* <http://www.scaltel.de/technologie-film.html> Verschaffen
> Sie sich einen schnellen Überblick über unsere Technologien,
> Dienstleistungen und Services.
>
>
>
> Erleben Sie SCALTEL jetzt auch bei <http://www.youtube.com/user/SCALTEL>
> und <https://www.xing.com/companies/scaltelag>
>
>
> - --To unsubscribe from enterasys, send email to lists...@unc.edu with
> the body: unsubscribe enterasys djo...@enterasys.com
>
>
---
To unsubscribe from enterasys, send email to lists...@unc.edu with the body:
unsubscribe enterasys way...@uca.edu
---
To unsubscribe from enterasys, send email to lists...@unc.edu with the body:
unsubscribe enterasys arch...@mail-archive.com
