Backing up your website

/

Ways to reduce website downtime

Distribute.IT hosting had a catastrophic event (or ‘epic fail’) recently when hackers who compromised its infrastructure irretrievably destroyed almost 5,000 websites. It is a timely reminder to be vigilant and strategic about backup and recovery.

As was just proven, even your host’s backup procedure may not be enough.  As many websites are often the primary shop-front and first point of contact for both marketing and sales, it is definitely worth review.

Websites and web applications are living breathing systems that in some cases have unique needs in terms of backup and recovery. There are a few points to consider when evaluating your level of required backup and availability.

1.    What type of website/application do you have?
2.    What is reasonable downtime?
3.    What are your customers’ expectations of your downtime?
4.     What is the cost to benefit balance between expected customer needs and site availability?

Before you say the obvious, that ‘downtime is not acceptable at all’, let’s look at some numbers.

Server Uptime Calculator

Availability Downtime / Year Downtime / Month (30 days) Downtime / Week
90% (one nine) 36.5 days 72 hours 16.8 mins
99% (two nines) 3.65 days 7.2 hours 1.68 mins
99.9% (three nines) 8.76 hours 43.2 mins 10.1 mins
99.99% (four nines) 52.56 mins 4.32 mins 1.01 mins
99.999% (five nines) 5.26 mins 25.9 secs 6.05 secs
99.9999% (six nines) 31.5 secs 2.59 secs 0.605 secs

Wikipedia [http://en.wikipedia.org/wiki/High_availability]

 

Each ‘nine’ could push your hosting cost up almost exponentially.

As at January 2010 Facebook’s uptime was 97.22% and LinkedIn’s was only 95.48%. That’s almost 16.5 days per year! [http://techcrunch.com/2010/01/04/twitter-uptime-facebook/]

So first, let’s mention domain names:

  • Own your own domain name with your active email address as the primary registrant contact. Don’t leave this in the hands of your web developer. If they leave town you won’t be able to renew it or change anything
  • Separate DNS control from your domain registrar and hosting company. Having this separation allows you to switch providers quickly in worst-case emergencies

Too much jargon? Apologies. Just ask your web guys to arrange it for you.

Hosting Scenarios

In our experience every client has different expectations, which usually differ from their budget. So let’s consider a couple of options for the budget conscious:

1.    Static business site
2.    More complex application or site with a blog

Scenario 1: Static Business website that is updated once or twice a year

Backup:

  • Ensure you get a copy of your site files from your web designer upon launch and after each update (unencrypted and not compiled) so that you could rebuild your site on another server with another provider if needed
  • Get a copy of the database that drives the backend of your site. Most websites have one
  • Some hosting providers offer an automated service that automatically transfers a copy of your site to a separate server of your choice on a weekly basis. (Cpanel offers this), however the mere fact of activating this service (or using cpanel for that matter) may open security holes that are unacceptable
  • Remember, most hosts don’t provide a backup service as standard, so be sure to ask

Availability:

  • If you are vigilant with the above points, your downtime could be extremely low – possibly only a few hours/year
  • Some SME’s may wish to go one step further and have their website files duplicated on another hosting provider, just sitting waiting and ready

This is a great insurance policy and considering that standard website hosting is only a few hundred dollars a year, a very cheap redundancy option. Just remember to choose a completely separate provider and geographical location.

Scenario 2: A website that runs an application that your business relies on such as a CRM or booking/payments system

In this case it all depends on your downtime expectations. Here are a few things you should check:

Backup:

  • The same rules apply as previously outlined with a few additions
  • Your database is critical – check what processes are in place here
  • Database servers usually (but not always) run a binary log that can be re-played to rebuild a database if it all hits the fan
  • Take point-in-time snap shots of your data and store the backups on separate servers (onsite and offsite)

Note: On basic hosting THIS PROCESS WON’T NECESSARILY BE HAPPENING – check with your provider but you may need a more custom plan and hosting solution.

  • Files on your web server will also need more regular back up, either across Raided drives or synchronized to alternative drives on and off the server
  • Still with me?

Availability:

  • Ask yourself what is acceptable downtime? Five minutes, four to eight hours or even two days?
  • Obviously, the shorter the time the higher the cost, however there are a number of Cloud hosting providers that can bring that cost down significantly compared to a few years ago
  • Traditionally, you would need to double every bit of hardware that could fail to reduce downtime and increase availability – load balancers, web servers and database servers, with the ability to switch between the failed components quickly

In the past this was very costly to implement and support.  However, Cloud hosting has provided SMEs an alternative, cost-effective option worth exploring. 

Cloud hosting automatically spreads your data and website across multiple hardware components, so it doesn’t matter if one component fails – your website will still be live. This effectively solves your availability issue, so all you have to worry about is backup! Yes, cloud servers and systems do have downtime, but the risks are considerably less.

Our Hosting Advice

As you can see, backup and availability quickly becomes complex with several options available depending on budget and existing infrastructure. Every business has different needs and expectations, so contact your web professional for advice.

  • Decide what is acceptable downtime and plan a strategy around that
  • What is your ‘worst case scenario’ recovery plan if everything fails?
  • Test your recovery plan
  • Plan your backup strategy between your hosting provider, web developer and you! Stay in the loop and test!

About the authors

Jason Hawkins is a senior application strategist, information architect and director at KND. His knowledge is based on years of industry experience in taking online application ideas to reality across a broad range of industries.

KND are web and digital professionals with over 10 years in the business. We specialise in web, cloud and mobile apps, eMarketing and SEO strategy, servers, hosting and design. We work closely with out clients to achieve successful outcomes. Call us today (+61 7 3832 4077) or email (enquiries@knd.com.au) to discuss your next project.

photo credit: Jamison Judd

Author:

Related Articles