Archive for the ‘Cloud Computing’ Category

Website/SaaS downtime: bad planning or bad luck

July 17, 2009

Websites, SaaS and web apps do suffer from downtime and slowdowns. These problems can result in a loss of revenues, customers and even reputation. Amazon looses $310,000 for every 10 minutes of downtime!!  Can downtime be avoided?  Yes, it can. The root cause of most  slowdowns or downtime is bad planning. But very occasionally it is bad luck.

Dilbert Downtime

The root cause of downtime: Plan to fail or fail to plan..

Just as a new building requires good design so does the software and hardware foundations of  a new website/web app. Both buildings and web systems also need regular maintenance or they will break.  Everyone knows this, right. Unfortunately these basic principles are sometimes forgotten during planning and once systems are operational. The result is unplanned downtime or slow downs. According to Yankee Group:

The major causes of Web site downtime are both technical – such as power outages, network failure, application problems and lax security – and human – such as inadequate staffing and monitoring or a lack of recovery planning.”

Causality (cause & effect) is at work here. Downtime is the effect and the  caused is often from one or several of the following:

  1. Poor code or app - Poorly written or design code causes downtime. It can also be a result of unsuitable platform or an app that’s been shoes horned into doing something it was not intended.
  2. Under capacity infrastructure - Hardware and software demands grow and change over time. One eye must always be kept on capacity and performance. Web infrastructure monitoring can help avoid this.
  3. Overloaded comms- This is probably the most difficult area to keep under control. You have to rely on a comms provider/ISP. So choose a good one. And don’t skip on the plan. If you do keep a eye on capacity. Website performance monitoring can help.
  4. Insufficient systems admin and management – Procedures and processes keep systems healthy and in tip top shape. Outsource it if you must. Again Website monitoring helps warn of growing problems.
  5. Bad luck - Multiple failures and failure of backup/contingency systems take websites down. At this point the disaster recovery plan should kick in. Unfortunately many people don’t have one of these.

The only point from the above list that can’t be avoided or planned for is Bad Luck. When website or web app revenues stakes are high why gamble with the chance of downtime or slowdowns. The blame for downtime often lies with poor management decisions. Corners are cut on resources. However as a business owner it’s a tough choice. Over architect and resources are wasted. Under architect capacity and your potentially setting up a problem for the future.

Related post:

The reality of website/SaaS downtime & slowdowns

The reality of website/SaaS downtime & slowdowns

July 3, 2009

As you can imagine the subject of website downtime/slowdowns is very close to my heart. Our startups new app monitors websites for uptime and performance. The full impact of Website downtime has just really hit me. This week part of The Pitch website frustratingly went down. On the same day the premier Web host provider Rackspace had a full hours outage which impacted many, many websites. These failures got me thinking. How common is downtime and what are the downstream effects for both ecommerce and SaaS apps/sites?

Downtime with the XXXX

Customers get angry when web pages are down. Don’t blame them.

Sh*t downtime Happens

As our good friend and Cloud Computing expert Simon Wardley says Cloud downtime is inevitable. Its going to happen. However that does not mean we should be complacent and accept it. Ecommerce consumers certainly aren’t. They’re increasingly less tolerant of downtime. The effects are real. A Harris survey in 2008 found that consumers abandon or switch 42% of all transactions when they have a problem. This is up 12% on 2007.

Also 84%  of online consumers share negatives  experience’s with others (Harris).  Many of these problems are technical and performance related. A reputation which took considerable effort to establish can vanish overnight. The downtime message spreads at the speed of light through Twitter across 10’s millions of users. You can become known for being unreliable.

I’ve done some routing around to find evidence of how common website downtime is. According to Pingdom the average website is down for two hours per month!! That could be alot of visitors or customers lost. Sainburys, a leading UK supermarket, website was down for three hours during one day last year and at various other times. A large number of customers were tracked immediately moving from Sainburys to a competitor.

Downtime costs

Back in 99′ Ebay’s 22-Hour outage was estimated to have cost them $2 Million. The Register figured UK  Website downtime to cost £565m a year in 2002. This included “indirect costs, such as loss of reputation, lost future sales and the cost of storing unsold goods”. Harris says (2008) transaction problems impact £11.9b of sales. The impact of downtime has other downsides. Google penalizes your long term Ads Quality Score when your site or landing page is down.

Despite the interruptions the outlook remains very strong for ecommerce and for other web services.  Cloud/SaaS is forecast to grow massively over the next five years. IDC are saying upto 33% of IT spending growth by 2013!! This growth includes many of the free services such as Gmail, Facebook, etc which we have come to love. As Cloud inertia grows more and more chargeable apps are going to switch to the web.

A very similar downtime and slowdown reality to ecommerce also applies software-as-service (SaaS) apps. Only the churn rate of customers is slower i.e. SaaS users/consumers are on short term billing contracts. The effect is not a immediate loss of sales. Unless they are trying to sign up of course. However the resulting loss of sales and reputation is the same. The Cloud and SaaS apps have much to learn from ecommerce’s 15 years of experience. And ecommerce still has a long way to go.

Related Post:

Website/SaaS downtime: bad planning or bad luck

Microsoft profits tumble: Big switch tipping point?

May 1, 2009

Is this the Tipping Point from on-premise software to software as a service? On the same week Microsoft announced a 32% drop in profits Apple reported sky rocketing profits.  Still the Redmond giant has a long way to fall. It makes a cool $1b Nett profit every single day! However, this is a striking contrast in fortunes during a recession. The downturn is acting as a catalyst of change and speeding up the Big Switch to cloud software services. The way we access and use technology applications is changing.

falling-stack-of-coins2At the heart of Apple’s iPhone success is the App Store and iTunes. Apple is successfully up-ending how music is distributed and how mobile phones are used. Microsoft’s poor financial results also follows the closing down of Microsoft Encarta last month. Microsoft gave up the fight because its traditional software could no longer compete with the likes of Wikipedia. The good and bad news keeps coming thick and fast. Only two months ago the SaaS CRM provider Salesforce.com reported a 34% increased in revenues and broke the philological $1b revenue barrier.

Its far too early to write Microsoft off but can the giant keep up with the increasing rate of change? Microsoft has rarely lead from the front but they have turned around and court up in the past. Microsoft’s response to the Cloud Computing Big Switch is Azure. The challenge for Microsoft will be to move away from its reliance on software licence revenues to a service based model. I’m not entirely sure they will be able to pull this one off..

Surfs-up startups: Where to be on the innovation wave?

April 15, 2009

A friend recently reminded me of Gartner’s Hype cycle. It got me thinking which part of the curve should  a Tech startup aim play in?

gartner-hype-cycle-july-2008Gartner 2008 Hype Cycle

Note Gartner  predominately focuses on the Enterprise business market. Although their right on the money in the consumer space with Microblogging i.e. Twitter’s rise to fame during the past year. The Gartner graph reminds me of  Geoffrey A. Moore‘s great startup book Cross in the Chasm and Rogers Technology Adoption Lifecycle graph. If your a tech startup and haven’t read it – read it now!

moores-chasmtechnology-adoption-lifecycleGeoffrey Moore’s Crossing the Chasm adapted Technology
 Adoption Lifecycle graph

I think startups should try to take advantage of Gartners ‘Visability’ curve growth.  PR opportunities can be maximised with everyone listening/talking about what’s new/what’s hot. However funds may run dry if a startup is completely reliant on the hype growth and doesn’t reach a healthy exit before the wave crests. Ideally, startups also need to be selling beyond Moore’s “Chasm” into the ‘Late Adaptors/Early & Late Majority’ which is where the higher revenues lie. All this is easier said than done.

Where’s your startup on the wave?

Cloud Computing: ‘Old wine in new bottles’?

April 9, 2009

I attended the Salesforce.com Cloudforce event in London this week and found this great introduction to Cloud Computing on the event website:

Mark Benioff of Salesforce is leading the charge against on-premise software. However the idea of Cloud Computing is not a new one. We had a false start earlier this century with ASP (Application Service Provider).  Before that Bureau services have been available for 45 years!  Today we have more acronyms in this market than you can shake a stick at – Cloud Computing aka Software-as-a-Service or on-demand services, etc.  All these names have the same underlying principle: a multi-tenant, pay as you go and elastic on-line service.

Is the latest title Cloud Computing a case of ‘Old wine in new bottles’? Perhaps it is. However technology has now enabled the old Bureau type services to be delivered with the rich GUI we have now come to expect from our PC’s and Mac’s. Cloud Computing – A new service 50 years in the making :)


Follow

Get every new post delivered to your Inbox.