I was discussing cloud computing with a client today, trying to find a good market analogy to position how cloud computing would evolve.
The electricity utility example doesn’t work. Neither does the water utility. There are two reasons: (1) Computing is a rapidly evolving technology, and (2) Service requirements vary widely for computing. Electricity production and distribution hasn’t evolved much since the invention of AC that made distance distribution possible. How many forms of water are needed around the world? It’s H2O – maybe it can be potable, purified, or come at a special temperature, but it’s still pretty basic stuff.
Analogies are never perfect, but done well, they can be illustrative. I think the best analogy for cloud computing services is transmitted music. Bear with me.
Music was first transmitted over radio waves 93 years ago. At the time, phonograph cylinders were used for music storage. There were probably pundits at the time who could look ahead and declare that individuals wouldn’t own their own copies of music – the radio would provide music for them “as a service.” Owning piles of phonograph cylinders, managing them – too expensive, too complex.
And in 93 years, we have gone from AM to FM to subscription-based satellite radio. It’s gotten better and better. The choices have grown substantially. The quality has improved dramatically. But so has on-premises music. The phonograph record, eight-track tapes (OK, not so great – but easier than lifting a needle to find a song), CDs, digitally-encoded music. Why hasn’t on-premises music died?
(1) Technology has continued to evolve both for off-premises and on-premises delivery of music. There is an arms race – while choices and quality grows off-premises, it also grows on-premises.
(2) Quality of service (and choice) matters. Yes, I can turn the tuning dial, and choose a different genre on Sirius, but if I want a specific song, now, and on a plane or in a cave or while I’m exercising, an Ipod delivers it. I’m willing to pay more to ensure I have choice and quality.
The bar for when off-premises vs. on-premises is the right answer will constantly change – there are many use cases where transmitted music services are good enough – and maybe better than I can afford myself. But there continue to be use cases where only on-premises can deliver.
And the same will be true with cloud computing services. Yes, a provider can aggregate users and drive tremendous economies of scale. Yes, managing your own IT is complex. But the same technologies that are driving and evolving to make cloud computing more automated and efficient are available (at a different scale) to end users. More importantly, service requirements can vary widely. Some services will naturally move to the cloud, and new services will evolve in the cloud, quickly. Other services will remain on-premises for quite some time (for differentiating, or service, or regulatory, or customization, or security, etc. reasons).
Will they eventually move to the cloud? I think that is very hard to answer, and maybe not important. The music analogy shows an example where the inevitable wasn’t quite so inevitable. The challenge is to find the right balance point, and try to predict how that balance point will move over time to ensure you get a good return on your on-premises investment. Predicting the next five years is important. And for the next five years, there will be many reasons to invest in private cloud services – with an eye on what’s coming in the cloud.