Results 1 to 7 of 7
  1. #1
    Join Date
    Jan 2014
    Posts
    36

    Cloud hosting uptime

    Hello, I'm new here and I want to check some things before deciding which service and which host to choose as my contract with current host will end next month.

    I've been having a lot of issues with downtime on my shared hosting plan this month. I'm not sure if it was only this month because I've been working at night on my sites the last week it might have come to my attention just now.

    I have counted already 12 hours of downtime this week. We have currently stepped into the web design business and suddenly uptime has become very important to us. I noticed this yesterday when my partner, who is in charge of sales, called me to ask why the site was not online whilst he was doing a sales pitch.

    As I'm in charge of all the technical stuff like the hosting and stuff, I need to get him informed ASAP on what he can expect from hosting. As we are in the process of deciding what we need I'm thinking of moving to Site5 shared hosting. But I'm afraid this will not be improving our uptime.

    So now i'm leaning to cloud hosting as it promises a much better uptime. But what does it really mean? When I calculated the uptime guarantee it actually makes no difference when you have 99,9% or 99,99% uptime guarantee. It still means 1 hour of downtime a month. Or am I doing the calculation wrong?

    What about maintenance? When there is maintenance does this mean another server will be taking over during maintenance of will I still have downtime?

    What about power failure or a disaster? If all servers in Amsterdam are out will Dallas take over for me?

    What can I promise to my partner? He is not really a technical guy so when I tell him cloud hosting will reduce downtime enormously without anymore explanation he will assume the sites will never be down. So I need numbers to explain him how much downtime he can expect.

    I know downtime can have a lot more issues than the server being down that are out of the hands of the host so preferably also theoretical numbers of that as well.

    Thank you in advance.

  2. #2
    Join Date
    Jan 2004
    Posts
    593
    If you are getting into the web development field and hosting those clients, then go with a solid provider with a great support team. Cloud or not, the reputation would be most important.

    Cloud shared and regular shared are really no different. It just takes one bad account to ruin it for everyone else. If you want the best uptime then you need to get off shared hosting. I'd recommend going with a managed cloud server. Those usually start around $50/mo. Spend the additional money and know your sites won't be affected by $2/mo shared hosting accounts. Its worth it and your clients will appreciate it as well.

  3. #3
    Join Date
    Nov 2005
    Location
    Nassau, The Bahamas
    Posts
    74
    Ditto to what speckl says!

    Spend some money, and get a great hosting partner. There's no such thing as 'cheap' AND 'good'.
    Secure Hosting | Premium Cloud & Dedicated hosting solutions since 2001
    email: sales(at)securehost(dot)com +1-242-502-8700 24/7 support by phone/helpdesk
    Located in The Bahamas & Bermuda: Redundant Cooling, Power & Network, PCI DSS compliant, 7ms from USA

  4. #4
    Join Date
    Jan 2011
    Location
    Canada
    Posts
    934
    99.9% is about 1 hr / month averaged. It doesn't mean you can't have 100% for a few months, long enough to boast about it and then have a multi hour downtime the next month.

    In reality, a single DC is hard pressed to push past three nines for any extended period. That's why bigger providers like Amazon AWS segment their service into availability zones. The goal of which is that if pushing past 3 nines is important to your business application, you figure out a way to span your application across AZs. These AZs are separate DCs across a city on separate power grids. A collection of AZs would in more general terms be called a region. Other providers also do regions. If you wish to guard against natural disasters, then spanning across regions is useful - but it's rare for it to be an honest business case since there are technical reasons that make it prohibitive to do.

    For those technical reasons, no cloud service will automatically failover to a far region for you. Some may do failover from hypervisor to hypervisor within the same DC. A rare few especially those on the VMWare side do offer AZ-level failover. But simple, straight-up cloud hosting will not reduce your downtime. Choose poorly, and it may even introduce a bigger single point of failure.

    In high availability, the real measure to look out for is time to recovery - not necessarily failure rate of your infrastructure. These two don't have to be bound together. But if your on a single VPS/machine, they are. Once you're able to split up into a cluster, then maintenance windows and hardware failures have very little impact on you.


    P.S. AWS also does AZ failover via their autoscaling groups.

  5. #5
    Join Date
    Nov 2005
    Location
    Nassau, The Bahamas
    Posts
    74
    tchen,

    Been in the business for 14 years, I disagree. :-)

    Any data center with N+1 infrastructure can achieve 99.9% uptime, easily.

    And in 2014, you should expect your cloud host to offer HA not only between hypervisors but also between near-proximity DC's, at least as an upgrade option. There's plenty of SAN to SAN replication options or software solutions like Veeam that make failover to another DC more common than you might think.
    Secure Hosting | Premium Cloud & Dedicated hosting solutions since 2001
    email: sales(at)securehost(dot)com +1-242-502-8700 24/7 support by phone/helpdesk
    Located in The Bahamas & Bermuda: Redundant Cooling, Power & Network, PCI DSS compliant, 7ms from USA

  6. #6
    Join Date
    Jan 2011
    Location
    Canada
    Posts
    934
    @SecureHostRichard The low number of Tier III (and above) certified data centres tell me otherwise. BTW, N+1 of active components is only enough to satisfy Tier II which is still rated below 99.8%

    I agree with the multi-DC rep though. But I don't think it's common enough (outside of enterprise) that it warrants repeating as an option/something different on WHT. Obviously the OP is new and wouldn't be blamed for believing that all cloud providers do this. The statement about Amsterdam-Dallas shows as much.

  7. #7
    Join Date
    Apr 2009
    Location
    New York City
    Posts
    5,054
    You should go with the most reliable provider you can find right now. Paying that extra fee or small price will assure that the hosting will be very stable and reliable .

Similar Threads

  1. Replies: 0
    Last Post: 12-10-2010, 01:08 PM
  2. Replies: 1
    Last Post: 12-02-2010, 12:00 PM
  3. Replies: 0
    Last Post: 11-15-2010, 04:05 PM
  4. Replies: 0
    Last Post: 11-02-2010, 01:39 PM
  5. Replies: 0
    Last Post: 10-13-2010, 01:08 PM

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •