We use cookies and other similar technologies (Cookies) to enhance your experience and to provide you with relevant content and ads. By using our website, you are agreeing to the use of Cookies. You can change your settings at any time. Cookie Policy.
Start a Discussion
Cloud Computing
user-pic

How Should Companies Prepare for When the Cloud Goes Down?

Vote 0 Votes
As this SearchCIO article, Time to lay down the cloud computing law for uptime, points out, it's when -- not if -- the cloud goes down. So how should you prepare for the inevitably of the cloud going down?

6 Replies

| Add a Reply
  • Simple, dont have mission critical services and applications running in the cloud. When it goes down, the impact is only a matter of frustration amongst staff, not customers....

    Obviously there are lots of technical ways in which you can prepare / handle such an event, however, I always like to keep things simple, and thats why I advise clients not to utilise the cloud for missino critical data / services...

  • The technology might be new but the problem is NOT.

    Business Continuity and Disaster Recovery (BC/DR) have always been risk management considerations... they were considerations before clouds, are still consdierations with clouds, and will remain considerations even after clouds.

    Wikipedia defines DR as "the process, policies and procedures related to preparing for recovery or continuation of technology infrastructure critical to an organization after a natural or human-induced disaster."

    DR is a subset of business continuity as BC involves planning for keeping all aspects of a business functioning in the midst of disruptive events, while DR focuses on the IT or technology systems that support business functions.

    So, in short, Clouds DO NOT eliminate the need for proper BC/DR planning and testing. Don't make the mistake of confusing or blurring the lines between a technology (cloud) and legitimate business issues (survival and profitability).

  • Although reputable cloud providers have many failsafe protocols and built-in redundancies in their architecture, outages are possible in the datacenter itself as well as in the Internet connection between the company and the datacenter.

    Even without data loss, which may be a costly result of such downtime, the disruption to business continuity is expensive as well. In my area of specialization, which is storage, we utilize a hybrid architecture that combines on-premises storage with cloud storage – that way even if the cloud goes down, local storage continues to function and in most cases, users will not even be aware of any issues.

    For applications operating in the cloud (i.e., SaaS), many vendors provide off-line capabilities that allow to continue to work locally and synchronize with the cloud application at a later time. This is an option companies should ensure exists for business-critical daily applications like CRM, email etc.

    Finally, of course it is impossible to guarantee 100% uptime, especially as regards the company’s internet connection, but building in redundancy is certainly one way of reducing the odds – for example, by having a backup internet connection running on a different type of infrastructure (DSL vs. Cable vs. optic).

  • I like this question. Drue says, "but every cloud provider goes down", Yes true, like any other systems, remote, local or whatever. "Most of them will pay for time lost during a service interruption, but not for valuable business lost, he said.": Is this worse than electricity power outage? I don't think so.

    As long as they are reasonably stable and cheaper than having local facilities, I use one of them. I don't want to see higher costs by asking them to cover too much liabilities.

    I will be implementing an ERP production environment for my company using a cloud computing service in this year. I make it sure that my data will be stored in at least two locations, either two remote locations, or one remote and one local.

    The worst case scenario in mind is rather that, someone physically cut off my optical fiber connection. I have to wait until it's replaced. While mean time, I may use mobile phone line, but it should be very slow than shared 1G bps, 100M bps bandwidth for me.

    Kengaku, Satoru
    S&S Global Services

  • This is a good trigger for reflecting upon service culture and contingencies.

    What do companies do when the electricity goes down or the phone is out? I am not familiar with outage statistics of the Internet at large or specific Cloud Services providers, but I venture to guess that their track record is not worse than that of the major utilities – probably even better. What is often missing is feedback from those service providers when they experience problems. I rarely saw providers that acknowledged a service interruption while that interruption took place. This is the most frustrating part – you do not know if the fault is within your sphere or if it is external.

    A very recent example is Orange in Switzerland – following the announcement of iPhone 4, their web site became overloaded and registered users could not access their account – unrelated to their interest in the iPhone offering. Yet, the only message you got when trying to log on was that it cannot present you a personal iPhone offer due to the high demand – and no word about the general login problem. It took about a week until you could log-in again. I expect providers to follow the example of Salesforce.com and be transparent about their on-going service level.

    Then comes the contingency aspect. Those who need constant electrical power install UPS systems. Those who need constant communications use multiple alternative networks. And those who need constant computing have DRP and facilities and processes in relation to their service tolerance. Why should using the Cloud be different?

    Now let me get back to the popular apprehension about the Cloud going down – and while we're at it what about the risk of your Cloud Provider going under? One of the most popular SaaS integration applications at Magic Software is the replication of Cloud based data – simply providing an integration link between a Cloud application (such as Salesforce.com) and a local DBMS hosted on the company premises. I consider this as some kind of life insurance policy – not too expensive, not a perfect solution, but something that would help you survive in case of the ultimate disaster.

    And putting things in proportion, data is probably safer and more available at Salesforce.com systems than in most companies’ data centres…

    My recommendation? Do your due diligence when choosing a Cloud service, require transparency from your Cloud provider in particular on the service state, and set up a contingency to cover your disaster tolerance.

  • Companies should prepare in the same way they prepare for when their resident equipment goes down. This means having good backup plans, disaster recover, and other tactics to ensure SLAs. At the end of the day a business needs to manage their IT services to well-defined SLAs. Can your organization handle 99.9% uptime for email? Maybe yes, maybe no. [Maybe people will welcome the break if its down :)].

    Everyone's mileage will vary, but there is nothing magical about the cloud. Its a pile of servers in someone's data center connected to your network pipe. Just like your company's data center only bigger. The "laws" of MTTF/MTTR all apply. Manage to the SLAs.

Add a Reply

Recently Commented On

Monthly Archives

Blogs

ADVERTISEMENT