Re: [OCLUG-Tech] Low Powered server

2014-05-02 Thread Spencer Cheng
Hi Peter,

On May 1, 2014, at 20:47, Peter Meyer petermeye...@gmail.com wrote:

 I ended up buying a Koolu (PC ION
 603) box with 512 Meg running on an AMD Geode processor.  This machine is
 still kicking and since that time I've added a USB hub and the data traffic
 to the attached drives has only grown.
 

I have one of those as well doing duties as a mail  web server. It’s a pretty 
wimpy box which can barely run spamassassin. 

 I am once again looking for a similar solution.  My new requirements are:
 
   1.  10 watts power consumption
   2. GigE
   3. USB3
   4. Sata3 (optional)
   5.  1G Memory
 
 I need something to run network data a bit faster than my current
 Koolu/USB2 Drive scheme.

The faster ARM cards can meet your requirement. I just received Nvidia’s Jetson 
dev card (Tegra 4). Haven’t had a chance to benchmark them but I would bet 
money that they are faster than the Tegra 3 based cards. The biggest weakness 
of 32 bit ARM CPUs is that their memory system tends to be slower (with few 
exceptions namely SOCs designed for server deployment) which means their I/O 
system tends to be slower.

You didn’t mention cost? :)

For general purpose server use, I would suggest one of the x86 boxen with a 
laptop chipset. They can come pretty close to 10W. I’ve seen a ultrabook that 
drew less than 10W under normal use.

/sc
___
Linux mailing list
Linux@lists.oclug.on.ca
http://oclug.on.ca/mailman/listinfo/linux


Re: [OCLUG-Tech] Linux Digest, Vol 113, Issue 3

2014-05-02 Thread Peter Meyer
Hi:

With regards to cost, I'd be looking for a solution under $300.

I've looked at the Intel NUC and other equivalent media boxes and the
idle power is above my 10W target.  The closest I've found comes in at 13W
idle is the Acer Veriton with the Centrino chipset.  I've confirmed their
specs using a Kill-a-watt meter.

I am leaning towards an x86 solution and appreciate the comments regarding
IO limitations with the ARM chipsets.  I won't be doing any more
complicated than:

   - print serving
   - DNS
   - DHCP
   - media share via nfs/afs/smb
   - file share via afs (time machine and data for video projects)
   - Squeezebox media server (yes its old, but still runs well)

Regards,

Peter


On Fri, May 2, 2014 at 12:00 PM, linux-requ...@lists.oclug.on.ca wrote:

 Send Linux mailing list submissions to
 linux@lists.oclug.on.ca

 To subscribe or unsubscribe via the World Wide Web, visit
 http://oclug.on.ca/mailman/listinfo/linux
 or, via email, send a message with subject or body 'help' to
 linux-requ...@lists.oclug.on.ca

 You can reach the person managing the list at
 linux-ow...@lists.oclug.on.ca

 When replying, please edit your Subject line so it is more specific
 than Re: Contents of Linux digest...

 Today's Topics:

1. Re: Low Powered server (Spencer Cheng)
2. Re: An email question (Alex Pilon)


 -- Forwarded message --
 From: Spencer Cheng sch...@aotera.org
 To: Peter Meyer petermeye...@gmail.com
 Cc: Linux@lists.oclug.on.ca
 Date: Fri, 2 May 2014 08:19:14 -0400
 Subject: Re: [OCLUG-Tech] Low Powered server
 Hi Peter,

 On May 1, 2014, at 20:47, Peter Meyer petermeye...@gmail.com wrote:

  I ended up buying a Koolu (PC ION
  603) box with 512 Meg running on an AMD Geode processor.  This machine is
  still kicking and since that time I've added a USB hub and the data
 traffic
  to the attached drives has only grown.
 

 I have one of those as well doing duties as a mail  web server. It’s a
 pretty wimpy box which can barely run spamassassin.

  I am once again looking for a similar solution.  My new requirements are:
 
1.  10 watts power consumption
2. GigE
3. USB3
4. Sata3 (optional)
5.  1G Memory
 
  I need something to run network data a bit faster than my current
  Koolu/USB2 Drive scheme.

 The faster ARM cards can meet your requirement. I just received Nvidia’s
 Jetson dev card (Tegra 4). Haven’t had a chance to benchmark them but I
 would bet money that they are faster than the Tegra 3 based cards. The
 biggest weakness of 32 bit ARM CPUs is that their memory system tends to be
 slower (with few exceptions namely SOCs designed for server deployment)
 which means their I/O system tends to be slower.

 You didn’t mention cost? :)

 For general purpose server use, I would suggest one of the x86 boxen with
 a laptop chipset. They can come pretty close to 10W. I’ve seen a ultrabook
 that drew less than 10W under normal use.

 /sc


 -- Forwarded message --
 From: Alex Pilon a...@alexpilon.ca
 To: David Patte ₯ dpa...@relativedata.com
 Cc: linux@lists.oclug.on.ca
 Date: Fri, 2 May 2014 10:35:20 -0400
 Subject: Re: [OCLUG-Tech] An email question
 On Fri, May 02, 2014 at 12:55:07AM -0400, David Patte ₯ wrote:
  Many websites currently use a persons email address and password as a
 way of
  insuring security for logging into a website.

 *Supposedly*.

  If someone has an existing email mailbox specified by a particular email
  address, can anyone on this list imagine any easy way that people can
  intercept email sent to that address without knowing the person's
 password
  at the mailbox site?

 Sure! Be root on the mail server and just peek at the mail spool.

 In all due seriousness though, that's very environment specific. I can't
 really provide a generic answer. Provided that basic security is in
 place, no, but you can't assume that it can be done easily. Still:

 * How many sites don't do TLS (and please, SMTP+STARTTLS and TLS ≥ 1.2
   ideally, not SMTPS)? IP traffic can and has been subverted en masse.
 * How many use export-grade ciphers?
 * How many accept certificates issued by untrusted CAs? Think of
   China and some businesses with man-in-the-middling firewalls.
 * How many end up having some relay in the clear at some point?
 * How many store emails on insecure storage?
 * How many are operated by untrusted administrators?

 And then there's poorly implemented SMTP [^1] servers.

 I'd look at how SMTP works first, if you want to understand.

 I recently advised a family member against providing credit card
 information over email sent to someone whose mail server only did SMTP
 in the clear. Scary how incompetent (and I do not use that word lightly,
 given the stakes) some mail server administrators are. There's missteps
 with little practical impact for what matters, then there's negligence.

 [^1]: Anybody still using UUCP, or have a need for LMTP?

 Regards,

 Alex Pilon

 

Re: [OCLUG-Tech] intercept email

2014-05-02 Thread Rick Leir

On 02/05/2014 12:55 AM, linux-requ...@lists.oclug.on.ca wrote:

If someone has an existing email mailbox specified by a particular email
address, can anyone on this list imagine any easy way that people can
intercept email sent to that address without knowing the person's
password at the mailbox site?
You have heard about the Pineapple? 
http://www.troyhunt.com/2013/04/the-beginners-guide-to-breaking-website.html


___
Linux mailing list
Linux@lists.oclug.on.ca
http://oclug.on.ca/mailman/listinfo/linux