[156957] in North American Network Operators' Group
RE: So what's the deal with 10Gbase-T
daemon@ATHENA.MIT.EDU (Tribble, Wesley)
Mon Oct 1 17:14:18 2012
From: "Tribble, Wesley" <WTribble@sterneagee.com>
To: 'Jima' <nanog@jima.tk>, "nanog@nanog.org" <nanog@nanog.org>
Date: Mon, 1 Oct 2012 16:12:33 -0500
In-Reply-To: <46422.2001:49f0:a057:0:f0da:c1ee:43f3:e40d.1349123576.squirrel@laughton.us>
Errors-To: nanog-bounces+nanog.discuss=bloom-picayune.mit.edu@nanog.org
10Gbase-T doesn't make much sense for a new virtual environment. Once you =
factor in the cost of the cabling and power, you probably would have been b=
etter off with DAC or FET interconnects. Also 10Gbase-T does not necessari=
ly work with Legacy wiring, depending upon how it was run. Large bundles o=
f wire cause crosstalk issues on legacy cabling, this is the reason for lar=
ge jackets on 6A. http://www.siemon.com/us/learning/alien-crosstalk-gui=
de.asp
I'm not saying it won't work for your scenario as I am not familiar with yo=
ur environment, just keep it in mind that with most environments, DAC is a =
cheaper and provides better latency for your storage traffic.
-----Original Message-----
From: Jima [mailto:nanog@jima.tk]
Sent: Monday, October 01, 2012 3:33 PM
To: nanog@nanog.org
Subject: Re: So what's the deal with 10Gbase-T
Gotcha. With SFP+ I think the only nod to backward compatibility would be=
1gbit RJ-45 SFPs, which can get a little spendy in large numbers (although=
so can DACs).
As for distance, I admit I haven't encountered any DACs longer than 15 met=
ers (~49 feet) -- not that I'm positive they don't exist.
Jima
On Mon, Oct 1, 2012 at 2:10pm, Andreas Echavez wrote:
> Mostly backwards compatibility; simplicity. We're planning for some
> super-high-density virtualization/storage projects mixed in with lower
> bandwidth gear, and sticking to one type of cable for everything would
> be convenient. I thought DAC had some distance limitations as well.
>
> This is all speculation though, I don't have any personal experience
> with the 10Gbase-T stuff either. I have no idea what to expect
> performance-wise.
>
> -A
>
> On Mon, Oct 1, 2012 at 12:58 PM, Jima <nanog@jima.tk> wrote:
>
>> > Does anyone here have experience running copper 10Gbase-T networks?
>> > It seems like the standard just died out. For us it would make a
>> > lot of
>> sense
>> > for our applications -- even if throughput and latency aren't as
>> great.
>> If
>> > anyone out there knows of any *copper* 10 gig-t switches (48
>> > port?),
>> I'd
>> > be
>> > interested to hear your experiences. I can't seem to find any
>> high-density
>> > ones from major vendors.
>>
>> Is there something unique about your environment that wouldn't allow
>> you to use 10gbit SFP+-based switches with DAC (Direct Attach Copper)
>> cables?
>> Those seem fairly well supported.
>>
>> Jima
>>
>>
>>
>
***************************************************************************=
***********************
Sterne Agee Group, Inc. and its subsidiaries request that you do not transm=
it orders
and instructions regarding your Sterne Agee account by e-mail. Transactiona=
l details
do not supersede normal trade confirmations or statements. The information =
contained
in this transmission is privileged and confidential. It is intended for the=
use of the
individual or entity named above. The information contained herein is based=
on sources
we believe reliable but is not considered all-inclusive. Opinions are our c=
urrent
opinions only and are subject to change without notice. Offerings are subje=
ct to prior
sale and/or change in price. Prices, quotes, rates and yields are subject t=
o change
without notice. Sterne Agee & Leach, Inc. member FINRA and SIPC, is a regis=
tered
broker-dealer subsidiary of Sterne Agee Group, Inc. Generally, investments =
are NOT
FDIC INSURED, NOT BANK GUARANTEED, and MAY LOSE VALUE. Please contact
your Financial Advisor with information regarding specific investments. St=
erne Agee
reserves the right to monitor all electronic correspondence.
***************************************************************************=
***********************
