Sun Project Blackbox / Portable Data Center

> Subject: Sun Project Blackbox / Portable Data Center
> 
> www.sun.com/blackbox
> 
>  
> 
> Has anyone seen one of these things in real life?

SLAC has a blackbox (which is actually white) 
installed, and running it packed with servers
for batch computing for the high energy physics program.

http://today.slac.stanford.edu/feature/2007/blackbox1.a
sp

Of course, using shipping containers for data centers
(and telco/networking) is not new, but this is a 
commercialized offering, rather than custom built
(although these early ones are still essentially
custom built).  

Note also that Google has (recently) patented
the "modular data center"

http://yro.slashdot.org/article.pl?sid=07/10
/09/1543256&from=rss

Gary 

On Sat, 2007-10-13 at 17:07 -0500, Lorell Hathcock wrote:
> -      Disaster Recovery

I can see portable generators being part of DR, but not one
or more
portable data centers.  How long would it take you to start
up a second
instance of all the hosts and devices you have in a data
centers?  Isn't
the purpose of DR to recover quickly?  I've seen a zillion
data centers,
and I've never seen two that look alike or carry the same
sub systems.
So the value of this idea is the case with the empty
rackspace
(IMHO).... but then I would have to pre-fill it with all my
same-kind
hardware and then store it somewhere safe until I needed it,
and I would
want it online so that I could keep it in sync... at that
point it's
only benefit is that I could move it from site to site as
hookup costs
(data/power) fluctuate.

> -      New Media / Web 2.0

HUH?

Like everyone else I think the idea is cool... just not sure
how
valuable it is.   Then again, CALEA brings a different
perspective, the
DOJ could have a thousand of these things on standby ready
to park
outside your offices when necessary. :rolleyes:

-Jim P.

Jerry Dixon wrote:

> We've looked at these from a DHS perspective and they
are a great
> concept.  I know Sun has had the boxes here in DC on
tour and worth
> checking out.  I believe FEMA was in process of looking
into leveraging
> them for disaster command centers along with the
military. 

As a long time network professional, volunteer firefighter,
CERT[*] team
member/instructor, and Red Cross disaster response
volunteer, I'd wonder
why ANYONE would want one of these.

If your "command center" is close enough to
require this specialized
configuration, YOU ARE TOO CLOSE.  Please stay somewhere
that you have
functional/reliable power, walls that are not falling down,
and hotel
rooms for your staff.

The idea of moving your data center INTO the hot zone would
scare the
bejeebers out of me.  I've been to large fires, hurricane
aftermath,
floods, tornado paths, and nowhere have I seen a need for
these things.

If you have a spare data center somewhere
"geographically diverse" from
your primary, you're golden and you don't need a unit like
this.

The concept of moving one of these things into an area that
is still
without normal utilities is not only dangerous to the
equipment "in the
box", but also begins to steal from the limited
resources that are being
used to rehabilitate the incident scene.

Why not get a couple of high-powered communication trucks
that would
allow network connectivity from the disaster zone instead of
endangering
your hardware and putting non-essential personnel into a
dangerous
situation?

Who's going to hire the private army that you are going to
need to
protect these monsters? (I understand that one is available)
And who's
bringing in the food for the poor technicians that are being
deployed
with the crate?

If you have the choice of using preciously scarce fuel to
power your
data center in a box or to help prepare and transport meals
to people in
need, which will you do?

Maybe I'm missing the point.

If you told me you were using one of these to deploy a
preliminary data
center while a permanent facility was being built, I'd think
you were
much more on track, but for emergencies?

AlanC

[*] Community Emergency Response Team, not computer
related:
http://www.wakecountyce
rt.org

Poor word choice on my part regarding command center versus
data  
service augmentation.

However there are many capabilities that this setup can
bring to bear  
no differently than a military TOC is established out in a
forward  
operating site.

I do agree that a good DR plan and hot/warm sites are a
necessity for  
critical services however there are uses for this
capability.  One  
only needs to look at Katrina or 9/11 where a solution like
this would  
have shown benefits in augmenting a larger DR plan and
quickly  
providing computing or network services depending on how it
was built  
out.

Just to add a disclaimer, these are my opinions and not an
official  
stance by the government.

Jerry
Jerryjdixon.com

On Oct 13, 2007, at 11:05 PM, Alan Clegg <alanclegg.com> wrote:

> Jerry Dixon wrote:
>
>> We've looked at these from a DHS perspective and
they are a great
>> concept.  I know Sun has had the boxes here in DC
on tour and worth
>> checking out.  I believe FEMA was in process of
looking into  
>> leveraging
>> them for disaster command centers along with the
military.
>
> As a long time network professional, volunteer
firefighter, CERT[*]  
> team
> member/instructor, and Red Cross disaster response
volunteer, I'd  
> wonder
> why ANYONE would want one of these.
>
> If your "command center" is close enough to
require this specialized
> configuration, YOU ARE TOO CLOSE.  Please stay
somewhere that you have
> functional/reliable power, walls that are not falling
down, and hotel
> rooms for your staff.
>
> The idea of moving your data center INTO the hot zone
would scare the
> bejeebers out of me.  I've been to large fires,
hurricane aftermath,
> floods, tornado paths, and nowhere have I seen a need
for these  
> things.
>
> If you have a spare data center somewhere
"geographically diverse"  
> from
> your primary, you're golden and you don't need a unit
like this.
>
> The concept of moving one of these things into an area
that is still
> without normal utilities is not only dangerous to the
equipment "in  
> the
> box", but also begins to steal from the limited
resources that are  
> being
> used to rehabilitate the incident scene.
>
> Why not get a couple of high-powered communication
trucks that would
> allow network connectivity from the disaster zone
instead of  
> endangering
> your hardware and putting non-essential personnel into
a dangerous
> situation?
>
> Who's going to hire the private army that you are going
to need to
> protect these monsters? (I understand that one is
available) And who's
> bringing in the food for the poor technicians that are
being deployed
> with the crate?
>
> If you have the choice of using preciously scarce fuel
to power your
> data center in a box or to help prepare and transport
meals to  
> people in
> need, which will you do?
>
> Maybe I'm missing the point.
>
> If you told me you were using one of these to deploy a
preliminary  
> data
> center while a permanent facility was being built, I'd
think you were
> much more on track, but for emergencies?
>
> AlanC
>
> [*] Community Emergency Response Team, not computer
related:
> http://www.wakecountyce
rt.org
>

On 14 Oct 2007, at 01:26, Jim Popovitch wrote:

>> -      New Media / Web 2.0
> HUH?

I understand what Lorell means - the web 2.0 scaling model
is to  
throw resources, rather than intelligence at your
bottlenecks.


I met some 'web 2' people at a conference quite recently,
and they  
were telling me their platform scales because they can keep
throwing  
servers at a cluster and performance increases.  Problem is
that it  
needs to scale early, and scale often.  I asked if any of
them  
understood the power requirements of a typical server,
whether they'd  
heard of the power constraints in the datacentres that
they'd all  
heard of, and how this model affected their new company's
OSS costs  
long-term, and none of them knew.

Scaling meant something else when I was solving these
problems for  
the first time.

On Sun, 14 Oct 2007, Andy Davidson wrote:
> I understand what Lorell means - the web 2.0 scaling
model is to
> throw resources, rather than intelligence at your
bottlenecks.

I think this is a little hard. Just about all the Web 2.0
presentations I
see have a big bit that says that how they had to redesign
and rearchitect
each time their customer level increased by a factor or 10
or so. The
newer companies are learning from this and implementing
scaling from the
start.

Most of these companies are less than a dozen people and
sometimes go from
nothing to "Top 1000" site in months or a year or
two. The aim these days
is to make sure you can do that.

Take a look at Pages 8-11 of this ( ppt -> flash
presentation):

http://s3.am
azonaws.com/slideshare/ssplayer.swf?id=122183&doc=aiders
s-aws-the-startup-project708

These people don't care about power, space, aircon and
bandwidth problems.
They just buy off others companies (eg Amazon) who
specialise in those
problems and charge the Web 2.0 companies what it costs them
to solve.

As for where the Blackboxes will be used, It'll be where
companies want
servers in place in weeks or months and existing datacenters
are full or
in the wrong place. Think of a building full of people
processing
insurance claims in India or a cluster delivering video on
demand in each
Asian city with more than 500,000 people.


-- 
Simon Lyall  |  Very Busy  |  Web: http://www.darkmere.gen.n
z/
"To stay awake all night adds a day to your life"
- Stilgar | eMT.