The Downside of Good Things

Recently I returned from a trip abroad and switched on my PC. After only 8 days away the number of components needing updating (Windows / Office / Norton / Adobe / HP / Safari / iTunes..) combined to break my personal record. It was 51 minutes from the time I switched on my PC to the time I was first able to do something productive.

It was only 2 years ago that I went past the 30 minute mark. It can’t be long before the hour is reached. Has anyone else got there before me?

While drinking my second coffee, I wondered if this was evidence that moving to a thin-client cloud-base would be more productive. Then I remembered last December when staying down in Dartmoor. The internet was running at dial-up speeds and fell over every 20 minutes or so for 4 days.

Looking at total system availability over the last year, I have had more time lost to network problems than to hardware or software problems. I am not alone. Computer Weekly reported recently that power outages and communications failures are the biggest causes of business disruption, contributing 3 times more incidents that hardware outages.

How much of a break will this have on the development of Cloud Computing? Adding an extra 9 to 99.99 availability at the data centre end is all well and good, but what is the benefit of that, if the weakest link is elsewhere.

One of the consequences of the visibility of blogging I have found is sometimes that I get approached by readers who would not have found me any other way.

The research department of a US law firm found my earlier blog entry on energy security. In the aftermath of the terrible devastation in Japan, they called to ask me my opinion on one question:

‘If the Japanese infrastructure was largely cloud-based would Japan have been more or less resilient in the aftermath of the earthquake and tsunami’.

As good systems people, we know that all systems to be effective need redundancy. We now have the sight that production of cars and electronic items are being hampered because of a shortage of components from Japan.

As a discipline, we in ICT have now contributed to making the global supply chains so efficient it is quite extraordinary, even looking back 10-20 years.

Yet, have we sought efficiency at the expense of resilience? Are we potentially creating the types of systemic risks that hit the financial sector in 2008?

Efficiency is a good thing. More efficiency must be better than less efficiency. However, reductio ad absurdum, we get some strange effects. Markets are efficient. Efficiency is a good thing; therefore each sheet of paper used by the civil service should be separately tendered for by competitive tender... to get the best deal for the tax payer. This is obvious nonsense.

Just when Japan needs economic activity to contribute to the recovery, that activity is being hampered by past efficiencies.

In Japan, some areas have lost both power and parts of the communication infrastructure.

The UK gets only minor tremors, but the flooding in Gloucestershire nearly hit the National Grid back in 2007. The Pitt review is interesting reading on that incident!

In responding to the American question, I came to a personal judgement. Within a decade, a major organisation may well go into administration because it cannot access its operational data in the cloud after an outage of power or telecoms.

So, the interesting challenge for public, private or hybrid clouds I suspect is ‘who manages the business risk?’

My concern is that IT cost efficiency through Cloud-based Computing badly handled may contribute to increased systemic business risk.

My experience of the Japanese is that they are that they are a proud people and they will rebuild. They will also learn the lessons of this disaster. They will adapt. While we in the West spend our energy trying to pin the blame on someone, they fix the systems.

For cloud computing to achieve its maximum potential, we need efficiency and resilience. The Service levels that matter are end to end and as seen by users. How will cloud suppliers deliver that?

I’m getting deeply unimpressed by the huge size of some updates and the productive time lost, but I’m not giving up my data yet!

So the question I want to pose is this:

‘For Cloud Computing to mature in the enterprise space, what availability of power and comms is needed before business risk is acceptable?’

If you can answer that, then when that will be deliverable?

Comments (3)

Leave Comment
  • 1
    John RUDKIN wrote on 16th Apr 2011

    Chris, you pose a really interesting issue. Responsibility in a fallible world....where does the Buck stop, and can it actually stop anywhere. The UK are worse than anyone for being risk averse, but the truth is, no one knows what will happen, just what could; and the permutations of things going wrong is staggering. If vendors of cloud solutions say 99.99% uptime, which part of their solution infrastructure do they mean.
    I've been playing with the Education offerings from Google and MS ( and have been a mobile me user for ages). It is surprising how that 0.01% can manifest itself. Who can think that twice in a month a street contractor in Blackpool would dig up our fibre. Who'd guess that the provider used a subcontractor to lay it in the first place, and that they decided ti lay it OVER hurried tram tracks?

    To cope, no solution should be 100% dependent.on. 99.99% solution. Instead ....... Swell I know what I'd do to mitigate the Risk, but what do others think?

    It's all very clouded really.

    Report Comment

  • 2
    jscole wrote on 3rd May 2011

    The problem eventually comes down to money. You can design a resilent system but if 50% of the cost is for a 0.01% probability, the bean counters will say that it is counter productive. The true cost of not having a contingency for the 0.01% probability happening is (as the Japanese disaster has demonstated) is much greater; but as a systems supplier, the counter argument is that that is the customer's risk, not the vendors. Whilst there is always going to be a vendor / customer division, we must work in a more 'joined up' manner, where the customer is willing to pay more to enable the vendor to protect the interests (financial, human, etc) of the customer. The problem is that the customer is not willing to pay for their own protection - they will only deal with the lowest cost provider (i.e. the one who has sacrifice resiliance) and sit fingers crossed hoping that the inevitable does not happen.

    Report Comment

  • 3
    Stephen Cole wrote on 14th May 2011

    Cloud computing makes no sense. Why develop high capacity and high speed secondary memory components (hard-drives, flash drives, etc), and then choke yourself on the network bandwidth?! The necessity for updates is mute as it is not immediately necessary to continue functioning. Updates should be a quiet time activity performed automatically by your computer. It suggests the need for more intelligence in the OS, not the need for a whole new computational framework like Cloud computing. The cloud is good for raw data and its sharing. It is not good for application download on demand. Realistically, the whole concept of the Cloud is purely about bean counting. Something I find total nonsense given that the action of counting the beans accumulatively costs more than just getting on with the job!

    The deep seated problem is more societal and not technical. The cloud computing concept as it is currently envisaged is a nonsense. I do see a cloud computational environment of some kind, but more along the lines of that developed by SETI, where economic number-crunching is needed. Otherwise we might as well just call the whole world a processor and turn it into a mega core silicon lump!

    Report Comment

Post a comment

About the author
Chris is a technology and policy futurologist. Chris has been in the IT industry since 1980. His roles have spanned Honeywell, ICL, HP, Microsoft and Capgemini. He is a Fellow of the BCS and a Fellow of the RSA.

See all posts by Chris Yapp

Search this blog

January 2018