Silverback Migration Solutions Data Center Blog

Hello everyone! I know many of you have been tracking “data center in a box” projects very closely. Projects like the Sun Blackbox and Rackable’s IceCube are ushering in a new trend in datacenter builds.

Several months ago I blogged about International Data Security, (IDS) and I have been working closely with Barry Prince from IDS during the last few months while IDS worked diligently to get their first ship based datacenter built and ready for customers.

Building a datacenter on a ship is not a new concept, it has been done on military vessels for decades. The military has already figured out how to secure the racks avoid damage from rolling seas, how to condition the air to removed the salt and added humidity, etc. All of these types of issues have known solutions, if you know who to ask.

Enter Vice Admiral Richard Naughton.

Admiral Naughton brings that knowledge to IDS as their President. He is well versed in outfitting large vessels with expensive, complex and often dangerous equipment. I was in a meeting with Rackable and IDS when someone asked about the affects of ship movement, and Admiral Naughton said something to the effect of, “When a destroyer gets hit by a submarine and jumps 6 feet out of the water, don’t you think I still want all of my systems still working. I need to know my systems are available to blow that SOB out of the water.”

When you put it that way, it does make sense the military would know how to build very resiliant systems. For them, it really is a matter of life and death.

The value proposition for ship based datacenters is very similar to that of land based datacenters, with a few noteable exceptions:

–Current market demand for data center space continues to outpace
supply, and using ships as data centers can reduce time to market by as
much as 65%.

–Cap-Ex costs to bring a ship into data center operation is
approximately 2/3 that of a land-based facility.

–Ships are naturally resistant from earthquakes and floods, and can
be moved to safer areas avoiding major hurricane or tornado storm damage.

–Vessel-based data centers can take advantage of their water-born
location to aid in cooling and thereby lower electrical costs. Locating
ships near electrical generation facilities lowers power loss over
transmission lines, maximizing efficiency. This makes ships a more “Green”
data center facility alternative.

–No in ground fuel tanks reduces environmental impact

–Recycling the heat from DC to heat the state rooms

–Emergency backup generators uses Bio-Diesel

–Recycle ships that would be scrapped

–More CPU cycles per kilo watt 65% or more of power goes to cpu
rather than facility

–These vessels will be located very close to major urban centers
where land cost and data center availability is problematic: customers can
quickly and efficiently update their processes and equipment without
traveling to remote areas.

–IDS marine data centers are the only data centers that provide
DRBC command centers with redundant communications, berthing and training
facilities.

–Process patent filed including 22 claims for a marine based data
center with all applicable trademarks having been filed both domestically
and internationally in EU and 7 Asian countries

IDS is currently in the final stages of due-diligence with their funding
and they claim the first vessel will be operational in late Q3 2008.

So, knowing that the ships are sound, with systems and infrastructure being built by the best in the industry, and being designed for life and death situations…why WOULDN’T you want to build your datacenter environment on a ship?

That is not a rhetorical question. I am looking for someone to shoot sound holes in this plan. If you have questions or concerns about IDS’ plans for ship based datacenters, please let me know by commenting on this blog or shooting me an e-mail directly at ken-at-teamsilverback-dot-com.

If you are excited about the opportunity of building your environment on an IDS dataship, shoot me an e-mail and I can assist you in getting the ball rolling to be one of the very first civilian customers with ship based infrastructure!

Last week, while taking a break from the general debauchery that is Interop, I had a chance to sit down with some old friends at Switch Communications.

If you haven’t had the pleasure of visiting NAP4, let me try to explain what a visit looks like…
First you pull up to a solid wall, surrounding what looks like a bomb shelter. You approach a speaker and announce yourself (and anyone else in your vehicle). Upon Switch Security verifying you belong there, they will open the gate and instruct you to drive around the building, inside the perimeter wall.

Once you find a parking spot, a door will open, and out will come a security guard that looks like a US Marine (because 2 weeks ago, he was probably still on active duty. Switch recruits military trained personnel for its security team). The security guard will very politely ask you for your government issued ID (drivers license). He will then verify it is you in the picture (and everyone else against their pictures) and will turn around and press a button on the wall.

Wall - “Switch Security”

Security Guard – “Mike Wiley with (names of everyone in group)”

Wall – “Have you confirmed the identities of everyone in your group?”

Security Guard – “Affirmative”

Wall – “Do you accept responsibility for everyone in your group?”

Security Guard – “Affirmative”

Wall – “Proceed”

Then the door will click and the security guard will open the door and invite you in to the first mantrap. He will then hand your ID’s to through the bullet proof glass to the guard (also an obvious Marine) while you look around a solid steel room with a revolving door. It takes your eyes a second to adjust to the dim room (not dark, just not Vegas sunlight). While you are busy feeling quite intimidated by the security protocols, the guard behind the glass is busy issuing visitor badges for you and your party. As each badge comes out, it is handed directly to the security guard in the room with you. He will then call the name of the intended recipient and will say “please wear this the entire time you are in the facility”. You comply.

Once everyone has a badge, the guard will use the military grade biometric scanner to allow himself through the revolving door. You wait until someone comes out to meet you, in our case Mr. Brian Boles.
Brian greeted us warmly and invited us into his facility. He scans himself to let each person in your party through the revolving door one at a time. The original security guard is waiting in the second mantrap for you to arrive. Once everyone has proceeded through the revolving door, Brian opened another door and we filed through, including the security guard.

Brian gave us a tour of the facility, including some environments Silverback has installed in that facility.
One thing to point out…everything in NAP4 is color coded…not just with sharpies and electrical tape. Powder Coat and High Gloss paint! Even the electrical boxes above the racks are painted red and blue is super shiny high gloss paint (more on that later).

During our tour, Brian took special care to emphasize the TSCIF above all of the racks and isle ways and mentioned the benefits of that patented design.

After our tour, Brian took us into an AWSOME conference room. I am a big fan of datacenter design, and anyone that builds their conference table out of ladder rack and brushed aluminum is pretty cool in my book. The overuse of ladder rack is phenomenal!
Brian then invited Melisa Young to the conference room to catch up with us and have a quick discussion about Switch’s plan for growth…SuperNAP.

SuperNAP is an UNBELIEVABLE concept. It is truly a Next Generation - State of the Art facility.

Let’s get the first, and most unbelievable part out of the way…1500 watts per square foot. That’s right, you read right. Not a typo…ONE THOUSAND FIVE HUNDRED WATTS PER SQUARE FOOT!

After we picked ourselves up off the ground, Melisa explained how they can manage that over 407,000 square feet. The concept is actually quite simple. Get the utility to provide more power to you than they provide to the Belagio, the Venitian, and Cesars Las Vegas combined, add a completely redesigned, patented cooling system (thanks Rob Roy), and System + System power and cooling equipment… … …shake and serve.

That brings up my next point. Switch has a 100% uptime guarantee on power. Not 99.999%…a ONE and TWO ZERO’S! ONE HUNDRED PERCENT UPTIME GUARANTEE ON POWER – FULL SLA…100% UPTIME! How on earth can they do that??!?! Easy…System + System power.

We are all familiar with redundant power supplies, circuits, and pdu’s…but Switch went further. They went to the UPS, then to the ATS, then to the building feed, then to the sub-station…then to the grid!! Switch’s power redundancy goes all the way out to separate power grids!

That means the utility company could lose a grid, and Switch customers still have power. Hell, BOTH grids could go down and Switch customers could lose power…Both grids, and an entire UPS could go bad…and Switch customers still have power…I could go on…and on…

Power distribution at Switch is broken into multiple Systems: Red, Blue, and Grey. Each device must be plugged into at least 2 distinct systems to qualify for the SLA. Single corded devices must use a dedicated ATS to qualify. Each part to the electrical system is color coded to the system it belongs to, right down to the conduit it runs through.

I know what you are asking…how do you cool 1500 watts per sqft in a datacenter when Leibert can only commit to 250 watts per sqft? Apparently, after Rob Roy heard he needed 200,000 sqft of Leibert devices to cool 407,000 sqft, Rob visited the Leibert factory and talked to them about getting more cooling out of less equipment. While I am not sure how that conversation went, I do know Rob set out to design a completely revolutionary design that has no cooling equipment on the raised floor. All of the cooling equipment is outside and above a false ceiling.

Enter TSCIF. Thermal SCIF (TSCIF) is another Rob Roy brain child. Every hot isle is contained and forces all heat up into the false ceiling, where is it sucked into the CRAC units. Pushing only hot air into the CRAC units makes them run much more efficiently, and reduces the overall temperature in the datacenter. Cold air is then pumped in through overhead ducts and is blown into cold isles.

When you factor in the security, redundancy, innovation, and dedication as 4 of the core factors in picking your next co-location or datacenter facility, Switch Communications wins every time.

Oh! I failed to mention the “other” neat thing about Switch…27 major ISPs on net…

Brilliant.

Here are some specs on SuperNAP:

- 407,000 square feet of space

- 7,000+ cabinets

- 250 MVA Switch owned substation

- 146 MVA of generator capacity

- 84 MVA of UPS supply

- 30,000 tons of system + system cooling

- 4,500,000 CFM

- 30 cooling towers

- 100% heat containment using thermal-scif

- Designed for 1500 watts per sq. ft. density

- Armed 24×7x365 military trained, Switch employed security staff

They have a pretty awesome video on their website outlining some of the features of SuperNAP, including a demonstration of TSCIF and how it works. Check it out here. (Make sure your sound is on…)

Silverback completed a full scale production migration for IMVU last October and we finally got the approval to take pictures of the environment from both IMVU and their data center facility!!

It is pretty hard to take pictures in a data center, security restrictions and all…but with the proper approvals, and a security officer standing over me and checking each picture as I took it, to ensure I didn’t take pictures of the wrong things, we were finally able to get some pictures we can share with the world.

Since IMVU has a fully redundant environment, Silverback was able to complete this migration with no service outage to IMVU’s core business.

All in all, IMVU was apparently impressed with the migration, as they now have an on-going contract with Silverback to install new infrastructure as required!!

Enjoy the pics!!

So, Silverback was chosen as an InteropNet Supporting Sponsor to assist with the rack, stack and cabling of the key infrastructure for the InteropNet network. Silverback is honored to have been chosen, as we have taken this opporunity to showcase our talents for the InteropNet team.

As we entered the Hot Stage Test Facility, we were greeted by 8 foot tall posters highlighting prior years InteropNet sponsors, some of them are even signed by the volunteers from that year.

That hallway leads into the break room, which is fully stocked with drinks and snacks.

Once you exit the break room and enter the main warehouse, there are shipping crates as far as you can see. Equipment was already arriving from this years sponsors, and the equipment from Interop New York 2007 was being un-crated and set out for the Interop Las Vegas 2008 show.

A quick look right reveals a mock up of the InteropNet NOC setup, including the sponsor tables, and all of the racks (referred to as “Peds”) that house the infrastructure for Interop network. 

We also got a glimpse into the exact WRONG way to run cables…don’t ever build a network like this…unless you absolutely have to, or as Geoff points out, “it will be torn down in 2 weeks”

Geoff Horn is the mastermind behind this project, and is often kept busy answering questions, and “making executive decisions (he) will regret later”.

When asked to describe InteropNet Hot Stage he said, “It can be best described as a controled plane crash, where about half way through, someone thinks we should put the landing gear down…does anyone know how to fly a plane?!”

To be fair, Geoff has the exciting task of building a complete, distinct, “new” network, using volunteer labor, and showcasing multiple vendors equipment.

Which he has become quite proficient at accomplishing. Although, in order to accomplish this monumental task, he needs his iTunes blasting Jazz or LoFi (whatever that is…) which competes comically with the OTHER stereo playing Good ‘ol Rock & Roll! Being in the middle of the warehouse was difficult and often resulted in singing Rock lyrics to Jazz instrumentals…

Ken Jamaca, Derek Baker and Cody Hays from Silverback have been working diligently to rack and cable much of the sponsors equipment as it becomes available. Here is Cody dressing cables for the “floor peds” that will feed the individual booths with internet access.

Cody eventually succummed to Geoff and Dereks commands to “do it like we tell you…not the right way!!” (Presidents note: I personally trained Cody to ALWAYS do it the right way, the first time…this must have been hard for him…) Using recycled cables, mounting gear in racks that roll away from you when you try to slide equipment in, stringing cables along the floor, not dressing cables and power cords as they go in the rack are all common InteropNet Hot Stage practices, which goes against all Silverback principles, although we understand why it is the way it is.

Every year InteropNet has a distinct network design, and as a reminder of this fact, they hang the prior years designs on the walls of the warehouse, they are a few years outdated, but they still have the full network topology from each event they have built.

…and if the network doesn’t work…we all know why…

Silverback Migration Solutions
is proud to announce…

MANAGED SERVICES AND OUTSOURCED IT

Silverback has assembled a team of highly trained, technically diverse and
proficient IT professionals to provide a complete arsenal of fully managed services
to all of our clients, worldwide.

Here is a short list of our new offerings:

• ‘Network Operations Center’(NOC) Services / Monitoring

• Customer Support Center (CSC) creation, staffing and management

Our technical staff members can quickly and efficiently augment your existing
team. If required, Silverback can even provide an entire outsourced IT solution
covered by our iron clad SLA structure! From process, procedure, monitoring,
documentation and labeling, to architecture, design, implementation and support;
Silverback Migration Solutions has the expertise to ensure your project is successful.

We understand your datacenter and IT infrastructure is the backbone of your
business and we understand how to keep that backbone in optimal condition while
you focus on growing your business.

Please feel free to contact our sales department at 925.407.4704
or silver-sales@teamsilverback.com
to discuss your current or future project requirements!

I was recently introduced to a new sales guy at XO Communications, and he is a phenomenal account rep.

He spent many years as a geek, IT Project Manager and WAN / IP Transit specialist; however, Jon has found his calling at XO Communications.

Jon Pauling is a Senior Account Executive for XO, and is currently helping customers set up MPLS WAN’s for voice and video.

I highly recommend Jon and XO Communications for any IP transit, long haul transport and private line connectivity. He is willing and able to provide the best services and solutions available today. Feel free to contact Jon directly at jonathan.pauling@xo.com.

Silverback Migration Solutions partner Servprise International has just released a Nagios plugin for all WebReboot products.

Nagios, ranked by eWeek as one of the Top 20 Most Important Open-Source Apps of All Time, provides Enterprise-grade network monitoring and alerting for all network-enabled devices and services.

The typical lifecycle for an IT failure involves:

1) A problem occurring

  2) Service degrades

  3) Users experience failures

  4) Lots of time passes until a user cares enough to report the failure

  5) Users report failures

  6) Lots of time time passes until admins receive failure reports

  7) Admins determine cause of problem and correct it

  Service is restored

When IT professionals are asked what it’s like to find out about problems  from users, all too often, the answer is “It makes us look like idiots.”

Now, with network and service monitoring in place with a solution, like Nagios,  the situation improves considerably. The lifecycle for an IT failure in this  case is more like:

1) A problem occurring

  2) Service degrades

  3) Monitoring system detects problem and service failure, and alerts admins

  4) Lots of time time passes until admins receive failure alerts

  5) Admins determine cause of problem and correct it

  6) Service is restored

The problem is, the users are still affected. The holy grail is to fix problems  before they affect users. To that end, what we’ve done with the WebReboot  Nagios plugin is tied the monitoring and alerting capabilities of Nagios together  with the ability to actually correct the problems, either through service level  or hardware level restarts. In this case, with this in place we get closer to our goal:

1) A problem occurring

  2) Monitoring system detects problem and service failure, and takes automated
  corrective action

  3) Service uninterrupted

  4) Admin gets to stay in bed. Users stay happy.

The current release of the plugin supports:

1) Automated service restarts when service failures are detected (think HTTP, SQL, or SSH, etc.)

  2) Automated server reboots when OS failures are detected (think “Blue Screen of Death”, kernel panic or segmentation fault, etc.)

  3) Automated power on when servers are accidentally shut down (think users doing “Start -> Shut Down” instead of “Start -> Log Off”
  by accident in Windows).

  4) Automated power off when temperature thresholds are exceeded (think someone tripped the breaker the HVAC unit is connected to on a Friday at 6pm).

  a. The WebReboot Enterprise reports on the temperature inside the server, and this information is collected and available for use within Nagios.

More details are available here:

http://www.servprise.com/products/nagios/NagiosPlugin.html

A PDF overview is available here:

http://www.servprise.com/pdf/WebReboot_Nagios_Plugin_Overview.pdf

Screenshots are available here:

http://dev.servprise.com/nagios/screenshots/

Cory Von Wallenstien, President of Servprise International can be reached at cory@servprise.com for more information, or if you prefer, you may contact Silverback Migration Solutions at silver-sales@teamsilverback.com

I spoke to my contacts at IDS and had some of my previous information updated…

• They are deploying 22 ships in North America - 50 worldwide.
• Only part of the 200,000 sq ft will be available in April, the remainder sometime around June.

I was also able to get the names of some key executives in the company:

• Founder - Ken Choi (linkedin profile for Ken Choi)
• President - Richard Naughton (linkedin profile for Richard Naughton)

They are working remotely until the first ship comes live in April. That ship will house the initial data center space, as well as the corporate headquarters for IDS. The first few ships in San Francisco will live at Pier 50. (I think they are there now, waiting to go into dry dock!!) 

Also, since these are “cargo ships” the whole “datacenter in a box” concept could get VERY interesting…

I have attached the marketing sheet for IDS to this post, the pictures at the top of the marketing sheet are of the first ship (going live in April 2008). IDS Datasheet

Silverback Migration Solutions is in the process of becoming a Sales Partner for IDS, so if you would like us to make an introduction, please feel free to contact us directly at silver-sales@teamsilverback.com!

Ken

Hello everyone!

I hope everyone had a safe and happy holiday season.

I would like to ask everyone for their professional opinions. Your response and opinion means a great deal to me.

I was introduced to a new datacenter company based in the San Francisco Bay Area. They are planning on building 22 new datacenters around the world. The interesting part is…they are building them on de-commissioned cargo ships!

The company is called IDS and they are building datacenters on refurbished cargo container ships. These ships will live in port in the their designated cities, and will rely on standard connectivity for power and network.

According to the people I spoke to, their approach lends them 3 distinct advantages over traditional brick and mortar buildings:

1.  They will use sea water for their chill water, reducing/eliminating the energy required to maintain appropriate datacenter temperatures.  This allows them to reduce their energy consumption by at approximately 30%. 

2.  In addition to standard datacenter backup generators, these cargo ships all have standard ship generators that can generate additional power. All generators will have access to the ships fuel storage during disaster scenarios, which allows them to operate for nearly a month without the need to refuel.  Most datacenters measure their backup power duration by days and hours before needing to refuel. Also, they can refuel via the water, which will allow them additional access to bio-diesel and less reliance on fuel companies access to surface streets.

3.  They are able to quickly provide large scale real estate in some of the most real estate impacted areas of the country.  They can outfit a ship and have it docked in San Francisco or New York in only a few months. Each ship reportedly has 200,000 sq ft of datacenter space!

They claim that they will have their first 200,000 sq. ft. of space available in San Francisco in the beginning of April 2008 and have already signed some key anchor tenants.

They have had excellent answers to all the questions I threw at them, so I am asking for your help; if any of you have heard anything about IDS, or have questions or concerns about their approach, I would love to hear from you. 

They are currently making their way out of stealth mode this month which makes it is very difficult to find information about them at this time. If any of you are interested in contacting them, let me know and I will be glad to make the introduction.

I look forward to your comments and responses.

Ken

This is the most interesting team building excersise I have seen…innovative…creative…entertaining…awe inspiring…

 More at their blog…