General discussion

  • Creator
    Topic
  • #2273244

    How close to build a redundant data center

    Locked

    by crackelb ·

    This is a disaster recovery question. I’d like to know if there are any industry rules of thumb to advise us on the distance at which we should place a permanent local redundant data center. We want it to be close for staff convenience, but how close is too close? How many miles would “typically” be evacuated for a “typical” environmental hazard, for instance? Does anyone know of any good resources to consult on this topic?

All Comments

  • Author
    Replies
    • #2725741

      What type of data center

      by thechas ·

      In reply to How close to build a redundant data center

      Is this just off-site data storage?
      Or, is the plan to move to this site for business continuity?

      For business continuity, you may need to be farther away than you might think.

      At a minimum, you want the second site to be served by a different power sub-station, if not a different high voltage main.
      Same for the phone and internet connections.

      Evacuation areas depend on many factors. Wind, specific chemical, available routes, etc.

      IMHO, the bare minimum is 5 miles away.
      Assuming the second location is up-wind of the primary site.

      Chas

      • #2720106

        Type of data center

        by crackelb ·

        In reply to What type of data center

        We intend to use this facility for storing live redundant mirrored applications, as well as failover infrastructure servers for active directory, email, telephony, mass storage, etc. We’ll have a redundant SAN and enterprise archive as well. Additionally, we plan to use it as recovery space for applications that need to be attached to SAN storage (for which the huge masses of data to be recovered would otherwise be prohibitive in the needed timeframe), or for those having high band-width requirements so must be on our fiber network. We’re a hospital – lots of radiology images. This facility may end up meeting any/all purposes including cold storage of equipment and backup tapes for more rapid recovery than we could achieve with storage with an off-site storage provider.

    • #2725693

      Cheaper to get a provider like Sunguard than build

      by jimhm ·

      In reply to How close to build a redundant data center

      Why build – when using something like SunGuard or one of the thousands of other DR providers do that for you …

      Thats a big waste of capital –

      I would recommend something like a Sunguard service or one of the other DR providers..

      • #2725628

        Consider Hazard Zones

        by curtis_n94521 ·

        In reply to Cheaper to get a provider like Sunguard than build

        Depending on what part of the country you are at, considering the area’s exposure to different types of disasters is probably more important than distance. For example, CA needs to worry about earthquakes, so a DR site anywhere is CA is not ideal, if this is where your primary business is also located. East Coast needs to consider hurricanes. Check with your insurance company who should have this info readily available. But it also depends on the size of your operation. Also consider JimHM’s post above. Not many folks create a DR site for DR’s sake anymore. However, you can have multiple primary sites that are designed to back each other up. In this way, you are not having a large capital investment that essentially sits idle.

        Good luck.

      • #2720103

        Why build?

        by crackelb ·

        In reply to Cheaper to get a provider like Sunguard than build

        Some of my explanation above about our plans for the center may explain why we’re considering a build vs. Sungard solution.

        We’re estimating this facility would need to be between 500-1200 square feet. IBM’s price for a local facility is $70/s.f. Local leased data center space is around $20/s.f. Sungard doesn’t have a local facility, and we need it to be on hospital fiber.

      • #2719999

        JimHM, think again

        by wordworker ·

        In reply to Cheaper to get a provider like Sunguard than build

        SunGard’s services aren’t cheap. I know the company well and like most of the people I’ve met who work for them, but they’re not by any stretch of the imagination going to be a less expensive solution for every company. At least if you build your own center, you have something left at the end of the day. With SunGard, it’s all insurance money – you pay it whether you use it or not.

        • #2718993

          Got to agree –

          by skipperusn ·

          In reply to JimHM, think again

          Sunguard is expensive for a dead center – and you got to wonder – what happens when two or three need the same center.

          One of DR drills failed because Sunguard didn’t have the facility. Now with HIPAA – the building needs to be more secure –

          So from his post about being on the Fiber – that isn’t as big an issue as HIPAA security over anything else.

          His concern about Fiber connection can be handled without much problem – even from a sunguard site.

          My recommendation is – list his major requirements – list his budget – is this going to be a dark site – then he could answer his own question… Or list them here and we don’t have to guess – and what his needs are…

        • #2713831

          More information needed!

          by damianthex ·

          In reply to JimHM, think again

          Before we can decide what’s the best solution, we have to know what the parameters are.
          How much down time is acceptible? What’s the geographical location (flood plain/tornado alley/etc.)? What’s the size of the data-center? How many bodies are available to the IT department?

          If down time allows and depending on the software you are running, a substitute data-center can be assembled in about a day with as many as twenty computers (with IT pulling an all-nighter) while adding more each day. Put a backup domain controller in your network off-site, buy a spare 50 port switch, a 1000′ box of cat5 cable, and a complete cabling kit. Keep these items with the BDC. Create a package that will allow you to install software FAST (one media per title, a list of software keys, the phone numbers to key software vendors, etc.) Do some research and find a local computer rental place that can guarantee the number of computers you will need to rent on a moment’s notice. Also find a location that will meet your needs that will be available at the time you need it, too. Air conditioned warehouse space should suffice for a large center, just use long folding tables. The tricky part is communication. It generally takes about a month to have a T-1 installed. You might want to get in touch with your provider to see what the best option is. Depending on the size of the center, you might consider having a satellite dish on hand to handle about twenty computers. It’ll be slow, but easy to install and it will give you Internet communication. This service will not cost a lot for a personal service account.

          The most important factor to consider when planning a contingency plan is to determine the mission critical nature of your data-center to the businesses survival. Would your business survive if offline for a week? Would two days put your company in the ground? The longer you would have to build an offsite center, the less expensive your plan will be.

        • #2714543

          ..definitely more planning needed.

          by ddt102 ·

          In reply to More information needed!

          In order to find how much downtime is acceptable.. It will depend on your business processes, their importance, and what applications and services you provide them. Perform a Business Impact Analysis (BIA) and define what is critical, necessary, and optional. Critical.. 4-24 hours, Necessary.. 24hrs to 3 days, Optional.. 3+ Days out..

          As far as the substitute datacenter.. It seriously depends on your size. Remember it is your backup.. your only safety net. If your datacenter doesn’t have proper facilities.. Power, Air Conditioning, Connectivity..(dist. to COs, availability of Fiber, etc..) your putting yourself at serious risk. The last thing you want to do is to throw together a DR DataCenter in an emergency situation. Just think that if you perform a DR to a warehouse.. and it has power issues or the Air Cond. systems haven’t been maintained.. You could be trying rebuild your infrastructure in an unstable environment.

          As far as renting computers.. If there is a significant disaster, everyone else in that vicinity will be also hitting up those computer rental vendors for computers.

          As far as DR.. network infrastructure.. It depends on your purchasing power with Vendors. Exelon can get a T1 installed within 3 days. But

          As far as performing “installs” with S/W.. I do recommend having a full copies of your software, patches, and customization at your DR site. However, what are you doing to do about customization of configuration, additional work by integrators/developers.. In the majority of cases, it is just not keeping a good inventory of your s/w, but having a complete DR/Backup strategy by Application. In some cases, it is easy to rebuild a box & install S/W on it.. in others.. you really want to perform a full recovery. If it is TOO complex to rebuild from scratch.. take the necessary time, work out the procedures, and perform tests..

          Again.. perform a BIA, review your architecture deeply, analyze the critical components and their recovery, setup a DR site in advance with possible matching hardware, store away system images (refresh often), configurations, at your DR site. Once you realize what systems and components are critical.. you then have the key requirements for your DC.

          Plan wisely
          -Drew

    • #2719998

      FEMA says no closer than 1.6 miles

      by wordworker ·

      In reply to How close to build a redundant data center

      FEMA recommends putting an alternate data center no closer than 1.6 miles from your main center.

      • #2713832

        Distances

        by electronic-service ·

        In reply to FEMA says no closer than 1.6 miles

        From experience being on a fire dept. we were working on a 25 mile distance for many years but then we started looking at a 50 mile distance as being better because the 25 mile distance could become borderline. I currently use the 50 mile distance for off site backup.

      • #2703760

        No closer than 1.6 miles

        by john.matteson ·

        In reply to FEMA says no closer than 1.6 miles

        However, you really need to take into consideration the environmental hazards you are trying to avoid, i.e. you are near a nuclear power plant. The evacuation radius for the power plant is 10 miles. If your main facility and your backup facility are inside that 10 mile zone, what good is your backup data center if both have to be evacuated because of an emergency?

        In coastal areas, where storms and flooding are environmental hazards, a mandatory evacuation is going to affect an area greater than 1.6 miles. Again, if both your facilities are inside the evacuation zone, what good is your backup facility?

    • #2719968

      Resources differ

      by miketalonnyc ·

      In reply to How close to build a redundant data center

      FEMA is a good place to start, as is any regulatory agency that governs your industry (i.e. FDIC for banks, SEC/NASD for securities firms, etc). There are a few good rules of thumb, however:

      1 – Never build a DR center on the same power/utility grid as the primary if you can avoid it. While the likelyhood that multiple grids will fail is small, though as August of last year proved it can indeed happen. Putting your DR center on another grid will at least lessen the chance you will lose all utilities at both sites.

      2 – go beyond the “line of sight horizon,” which will be different depending on where you are. Generally, at least three miles is a good start, 10 would be better.

      3 – try not to put both datacenters in “target zones.” As horrible as this is going to sound, major industrial areas are targets for lunatics of all kinds. Having a primary datacenter in New York City and a DR facility in Los Angeles or Chicago – for example – is not the best of ideas. While it’s true that finding the right type of facilities in non-urban-centers might be difficult, it’s still a good idea to place the DR facility someplace at least marginally safer than your primary.

      Hope that helps.

      Mike Talon
      miketalonnyc@yahoo.com

      • #2719907

        Other tips

        by miketalonnyc ·

        In reply to Resources differ

        Hi folks, I completely forgot to mention that TechRepublic.com offers a weekly newsletter on Disaster Recovery issues (actually they offer newsletters on many topics). You should be able to sign up with your existing TR.com logons, so if you’d like to see my opinions on DR issues each week, feel free to sign up.

        Mike Talon

    • #2701753

      Depends on what’s comfortable…

      by haze_kalasa ·

      In reply to How close to build a redundant data center

      I think you’d be very hard pressed to find a “one size fits all” answer to this one. There are alot of factors that come into play that are very unique to every location. What you have to ask yourself is what are you comfortable with. There are natural events (i.e., hurricanes, earthquakes, fires, or even a tornado) that can span very large areas of geography. Then you have the man-made eventts (i.e., building fire, bombs, etc.) that can be much smaller in scale. Basically you just have to play the odds and see what your organization is willing to define as acceptable.

      We have a cold site that is over 1,000 miles away for our critical applications. As the liklihood of both locations (our live and cold site) being hit with a disaster are extremely remote. We looked at a site that was only 20 miles away but being in the midwest we had to take the possibility of tornados into account. While the odds are very low of two sites, 20 miles apart, being by a tornado are very low it wasn’t something we were willing to risk. Yes there are some big inconveniences associated with having a site 1,000 miles away but when taken in conjunction with the potential cost of losing two sites it’s worth it.

    • #2701243

      Distance is related to structure

      by wrlang ·

      In reply to How close to build a redundant data center

      Everyones suggestions are great.
      I would add to the suggestions that if you build your backup facility using a fortress approach, it can be within easy commute distance. Fortress approach means that the new facility cannot be harmed by ANY of the threats you need to be concerned about for your area. Here in Wisconsin that’s Tornado, flood, fire, winter/summer storm, and terrorist (anthrax, explosion..), hazmat PROOF. Can be an expensive proposition, but…

    • #2713873

      How close to build a redundant data center

      by richabel ·

      In reply to How close to build a redundant data center

      One it is cheaper to go with a company that can supply you with a site for disater recovery. the differences here are what typpe of site you are going to have –

      a) Hot site – this where your current operating systems and application engines are loaded and all you have to do is use your data recovery backup devices (tapes or what ever your company uses currently).

      b) Warm site – this is where your operating systems are loaded and you have to put your application engines and data recovery backup devices (tapes or what ever your company uses currently) to get your data to currency.

      c) Cold site – Here you have to load everything

      The costs vary with the type of site. The one thing to keep in mind however is how long can you afford to be out business…
      As a good rule of thumb for disatnce the recovery site is one that is far enough away that so as not to be caught in the disaster situation as your “operational” site.

    • #2713864

      Some minor hints

      by gyrfalcon888 ·

      In reply to How close to build a redundant data center

      Do an Internet search for “ebook.pdf” and read it carefully.
      Your options will be constrained by how much of a budget you have and what kind of facilities you have to work with in the area.
      One option is to make a reciprocal agreement with a business nearby that has a similar technology environment. You could provide alternates for each other with some expansion.
      Whatever you end up with make sure that you do a live test to make sure that it works.
      However, if you have the funding to build a permanent local redundant data centre then you could probably afford to hire a proven expert to do this work for you.

      • #2713858

        Test at least twice per year

        by ghaney ·

        In reply to Some minor hints

        I agree with a lot of the previous comments. No closer than 10mi, no further than 4 hours away. You will also want to ensure that there is adequate office space at the DR/BC site for staffing needs, should it arise.

        One of the things we are looking at doing (depending on funding) is creating a hot-site with data replication. Yes, this is expensive; esp. when you look at the requirements for diagnostic imaging and the length of time required for data retention in a healthcare environment.

        We intend on switching production from our on-campus computing center to our DR/BC center in six month intervals. This will provide us with two live tests of failover per year and will ensure that staff is well-trained in how to switch operations from one location to the other when needed. It also allows for hardware at each site to be upgraded without impacting production (hopefully).

        Fyi.

    • #2713814

      What Possible Disasters?

      by henry_bg ·

      In reply to How close to build a redundant data center

      One of the first things you have to take into consideration is the location of the primary site. Let’s say the primary site is in Miami, Florida. 50 Miles north of Miami is still prone to Hurricanes so you may consider a location that is not suceptible to hurricanes. You have to choose a location that is not prone to extreme weather conditions such as flooding, tornados or excessive snow. Nowadays it may even be necessary to choose a location that is not likely to become a haven for a terror attack.

    • #2713799

      How far

      by cpr ·

      In reply to How close to build a redundant data center

      What type of hazard are you preparing for?

      If an airplane filled with bacteria crashes into a tall building, then 10+ miles would be adequate, if a dirty atomic bomb is detonated above ground, then 20+ miles.

      If you are going to setup a backup data centre with the current state of network communication, then basically any distance is acceptable. If you want your staff to get to this data centre, will they be able to leave their homes, can they use the airport, will roads be closed, traffic congestion, chaos, etc.

      Perhaps the best solution is to setup a remote center (100+ miles away), staffed by non-company personnel, and if a hazard occurs, somehow get key staff members out to this site.

      Remember, if a widespread hazard occurs, you will not be the only company affected.

      Regards

    • #2713762

      10 kt

      by macgold ·

      In reply to How close to build a redundant data center

      Depends on the threat you see and resources you have. 9/11 hit WTC, but paralyzed all NYC and northern NJ. Wonder how many NYC firms thought NJ was safe, until the latest terror threat named Newark, NJ, as well as NYC. It also depends on your resources. Worked for one firm with the primary data center 250 miles from corp HQ in northern NJ, another with data center 1500 miles away. If you have multiple sites, consider using them as secondary sites, esp if one is someplace odd i.e. an unlikely target. Thus, backup the NYC office to the LA office, and LA to NYC, or better, both to Kalamazoo.

    • #2714685

      Another good resource

      by frierog ·

      In reply to How close to build a redundant data center

      The Disaster Recovery Journal is a magazine that comes out quarterly and is an excellent resource.
      Check out their Website at http://www.drj.com

    • #2714653

      Who are your Auditors

      by johnboehlke ·

      In reply to How close to build a redundant data center

      Up until a short time ago, I worked for a health care plan. When we did or Business Continuity Plan we consulted with the Ernst & Young eRisk group. E&Y happened to be our auditiors but secondarily they do a lot of work in the health care industry. They recommended about 5 mi seperation.

      You might try to consult them or your own auditors for their view on best practices.

    • #2714600

      Think power, phone & transportation grids

      by oien ·

      In reply to How close to build a redundant data center

      Before locating a redundant data center I’d start with talking to your county emergency services coordinator. They should have maps of how things are divided for your area. The first item I’d look for is how is the power grid for your area set up. Is there a clear dividing line where a major substation outage wouldn’t cause a power outage in both facilities? Secondly, make sure the telephone and related data lines, if possible, come a different supplier than your regular service or at least have alternate paths to keep them alive. You know you have to be able to get your people to the redundant facility, but take into consideration the county emergency transportation routing system so you know the route to your facility will be accessible.

      Greg Oien

    • #2714522

      Why reinvent the wheel?

      by mjost ·

      In reply to How close to build a redundant data center

      The government has some great publications on the subject. Check out: http://csrc.nist.gov/publications/nistpubs/800-34/sp800-34.pdf

    • #2714394

      There is no easy answer to this…….

      by sleepin’dawg ·

      In reply to How close to build a redundant data center

      You can do your best to plan for every eventuality but as sure as God made little green apples some thing will happen to screw up the best laid plans of mice and/or men. Jan/1998, large portions of eastern Canada and Northeastern USA were hit by freezing rain which caused a literal collapse of the high tension power transmission system with it went a large portion of telecommunications especially cellular. It was over 30 days before the complete restoration of services although 70% was restored within 15-20 days. Even hurricane disaster recovery plans couldn’t cope with this type of disaster. It’s one thing to replace broken poles and towers it’s a completely different thing when you have over a thousand miles of towers down and have to chop through 12 inches of ice to fix them. While the majority of the system was restored in 20 or so days the complete repair and replacement took over six months along with the attendant stoppages. We thought we had worst case scenario covered since we have offices in Boston and Houston both of which get their share of hurricane activity but we could never have envisioned anything like this. Fortunately we had a disaster plan in place but not with the head office in mind. Head office was out for 9 days but our operations were only out for 7-9 hours and that only the head office portion. Are we ready for the next one??? We hope so but who can predict tomorrow???? Maybe next time it’ll be earthquakes, floods or worse, solar flares. Just don’t have all your eggs in one basket.

    • #2714392

      ARMA says 100 miles

      by sprodriguez2001 ·

      In reply to How close to build a redundant data center

      According to ARMA you should have backups stored no closer than 100 miles from the site where records originate. ARMA is a national institution for information management. For further information on this topic you can refer to http://www.arma.org.

    • #2713611

      Close as you can, with different services

      by delbertpgh ·

      In reply to How close to build a redundant data center

      The closer you are to it, the better you can manage it. But choose the place so that you are insulated from the disaster that took out your first center.

      If your primary site could be flooded, don’t locate your secondary on the shore of the same river. Figure out what kind of disaster (like flood, fire, riot, terror attack) you think you need to protect against. Then make sure your secondary site will not be blitzed by the same event, and won’t be isolated from your recovery team by the disaster.

      Get far enough away that your secondary is served by different electric power and telephone utilities.

      If you’re just locating a spare server in somebody’s spare bedroom and calling that your disaster site, these rules may be good enough. If you’re actually putting together something that costs a lot more, get a consultant. If you are trying to construct a high-reliability site to support a financial institution, go to Sungard.

    • #2714368

      How far for backup site

      by basisdba35 ·

      In reply to How close to build a redundant data center

      I would evaluate what area you are in, areas such as California with it’s earthquake faultlines, would suggest places that would be 50 to 100 miles or more away. Questions to ask:

      1. How much physical access do we need to the recovery site?
      2. Will users have an alternate site to work from? home, alternate office space, etc.
      3. What is the RTO in hours, days?
      4. While data can be replicated via different venfor products the hardware needed in the remote site to bring up the applications must be considered. How many servers are needed for business critical systems.

      I have done a good deal of work on the Business Recovery requirements and this is a much broader question than just IT recovery.

      • #2714278

        It’s dependant on number of company sites with IS/IT capabilities and dis

        by sleepin’dawg ·

        In reply to How far for backup site

        In the 1998 ice storm the complete electric grid was down in eastern Ontario, most of the indutrial
        sector of Quebec, large parts of New Brunswick and Nova Scotia, upstate New York, Vermont and New Hampshire, as well as parts of Maine, Pennsylvania and Massachussetts. Roads were closed for over the better part of a week, with a lot of places for longer. Air transportation was grounded for a week and local mass transit was out for up to two weeks. Even when the grid was brought back on line it was not reliable with lots of small intermitent outages over a 30 day period. A lot of the phone system went out for several weeks and cellular phones became next to useless as battteries could not be recharged. Standby emergency generators ran out of fuel and fuel couldn’t be delivered because either there was no power to drive the pumps to fill the trucks and even if there had been the roads were impassable for delivery. People couldn’t move even when roads were opened because a lot of bridges remained closed because of large masses of ice falling from their superstructures and even city streets were hazardous due to falling ice from buildings. Local hotels were booked solid with people who couldn’t get home and those people had to take up the slack for those at home who couldn’t reach work. We were lucky in that our head office was only fully down for 3 days, operating with about 40% of staff for close to 3 weeks but we didn’t get up to full staff for over 30 days and even then travel times to and from work necessitated shortened hours for many of our people. Most large buildings had backup power that was strained to capacity and many smaller buildings incurred damage from frozen pipes. Supplies of polyfilm ran short as many offices rushed out to buy rolls of it as a preventative measure. It comes in 4mil or 6mil, black or clear, in rolls of 12 Ft X 50 Yds. Clear was the preferred choice and offices that were stuck using black were very dreary places indeed. Surprisingly the morale of people was very high with only a few worry warts as to be expected and they were usually shamed into silence by somebody’s worse disater story. The area covered ran from Buffalo to Portland and from 50 miles north of Montreal to Boston and for 8-10 hours NYC. Our Boston facility was only down for a day but ran only 75% staffing for up to 10 days. On the basis of loss of life it was a minor disaster but on the basis of crippled businesses it was major. Where would I put a DRC??? I haven’t the foggiest because what good would it be if nobody could get to it and the only reason to get to it would be to service clients who are undergoing the very same problems you are. They wouldn’t have needed us until the situation was stabilized. Our solution has been to equip every PC with a UPC and scheduled backups to mirror servers at each of our offices (now 7). Each office has independant backup propane generating capacity which is tested monthly and of course all programs and OS’s are available including encryted copies of propietory IS/IT. We are a bit different than most places because we are a design, implementation and integration house and as one of my partners loves to say some of our work is right out of this world referring to some of the stuff we’ve done for NASA and the CSA. We even design hardware so we do not have any great need for a dedicated IT group. Total down time for specific offices was high especially Montreal but company wide we were hardly affected by more than 6 hours. Montreal is our biggest house with 63 full time people but the other shops are much smaller with 20-35 full timers. We do have out side contracts with approximately another 250-400 as the situation or projects demand. Even our contract people backup their work for us to one of our servers on a daily basis and these inturn our mirrored at each of our sights. It’s not cheap
        but on a cost effectiveness basis it’s cheaper than any of the services and it’s always available to us as needed and need is our fundamental priority. Plus we have equity in our own equipment and once the initial cost is absorbed the major expense is maintaining the batteries for the UPC’s,something we are told have a 4-5 year life but we turn then over every 3 years. DR software is reviewed on a monthly basis and upgraded as required something that consumes about 30 minutes of someones time whitch is cheap insurance for a peaceful nights sleep if and when the temperatures warm in January and the rains begin to fall for ten days straight. The measured ice thickness was supposedly 14 inches and I can testify that my cars sitting in my driveway were in ice to over the hubs of there wheels. In case you wonder how you dig a car out of that, you chip ice carefully around the doors and get a door open enough to slide in and start the motor. You will pobably have to chop ice on the ground to let the door open enough to get in. Do not use a remote starter as the motor shuts off after 10-15 minutes and if the battery is in top shape that might not be enough to recharge it fully after the next few starts. Turn on your defrosters full blast and the get out of your car and start chopping away enough ice to slide a jack under your car. The kind you can slide under the differential housing is best. Jack the car up as high as you can get it and very carefully slide the car on the jack sideways. It doesn’t have to be far 6″-8″ should do it. You can get additional height by putting one or two pieces of wood on the jack plate and its best if you slide the jack in from the side. Get a neighbor to help push and do it gently as you don’t want the car to slip off the jack and damage the under side of your car. If you have the standard screw type jack place it beneath the center door post again with a piece of wood, jack the car as high as you can and then push the car off the jack. The screw jack has an advantage in that it usually frees both ends of the car at the same time whereas, the other method is going to take longer but if anything goes wrong the repair is going to be more expensive. Any damage you get from the jack under the center post is going to be down low on the rocker panel where you’ll hardly notice it. Of course the real smart thing is to move the car a foot or two every few hours so it doesn’t get so iced in, in the first place. I did that with my car but was too dumb to do it with my wife’s car. What the hell you can’t think of every thing!!! She could have moved her own damn car. Six years later and I’m still hearing about it.

    • #2705818

      Resources and Job Market

      by ljhunt ·

      In reply to How close to build a redundant data center

      The only place, multiple communication company and power source junctions (the corners of the Grid).

      Where completely seperate communication services will complete for the business, empty corporate facilities are prime for the picking. Human resources are readily available or would not mind moving into the are.

      Secret — I’m working at that place now, due to lease rates and location. We are now the Primary site and the original site is noe the redudant location.

    • #2722033

      Weather?

      by marco schumacher (at biznesslegion) ·

      In reply to How close to build a redundant data center

      Consider how important weather might be, both as a vulnerability for the data centers and a hindrance to the staff that need to get there. For example, can you afford to have both locations in a zone with heavy snow fall, which may break power lines and hinder traffic? Should both be in the likely path of a hurricane?

      (Similar considerations apply for call centers. Backup centers for US East coast locations are often in the Southwest. This provides time zone benefits [easier on the employees if you need to have 24 hour coverage] in addition to the weather advantages.)

Viewing 20 reply threads