Off-peak Computing – Economy 7 for the Cloud?
My dad’s house is full of power outlet timer switches, which turn on the electricity supply during the night when the utility company provides cheap off-peak electricity. The washing machine, dishwasher and even the bread-maker are all set to run at night. In the UK, this is known as an Economy 7 tariff and in other regions it is called “time of use tariff”, or simply – “off-peak rates”.
Why is this relevant? Well other than merely pointing out that my dad loves a good deal, the topic came up when I met a customer last week, who asked if I could offer cheaper rates for computing resources that are used during the night.
At the time, we shared a joke about “Economy 7 for Cloud Computing”, but there is a serious point; computing is becoming more synonymous with the economies of the energy market. But how far does this analogy go and how far should it?
Utility pricing in the world of Cloud
Cloud Computing is all about utility based pricing; paying for only what you use with no upfront capex costs and unlimited capacity. This is just the same as the energy market and it is a nice analogy, so no wonder people are comparing the two in order to better understand the cloud approach for compute power charging.
The comparison gets somewhat confused when it is extrapolated into the realms of off-peak rates. At a local level, for local resources, there is logic in this concept. It’s possible to avoid under-usage while scaling for the inevitable peaks in demand by incentivising usage outside of peak hours. An Australian data centre, for instance, may have more demand during their standard working hours than at night, therefore encouraging higher usage at night is economical.
However, public cloud computing isn’t regional, it is global. If this Australian data centre did offer cheap rates for night time consumption, then it would make economic sense for European organisations to sign up and take advantage of this cheap compute power when they need it most; their day time.
In this digital age where connectivity is crucial and data is king, global businesses are integrating and connecting their customer base around the globe and their IT demands inevitably follow suit; they become global. As such, regional computing discounts only make sense when data residency requirements come into play and a regional specific or private cloud is used, meaning that there truly is a need to “flatten” the demand profile.
Compare the (Cloud) Market
Moving past the off-peak analogy, there are other systems in the energy market that also have relevance as cloud computing matures. For example, like my dad, I also love a good deal. When I buy my energy, I want the cheapest rates so that I minimise my costs.
To do this, I make use of price comparison sites like moneysupermarket.com. I plug in my requirements and typical energy usage and out pops the results, which are ranked with the cheapest at the top. I do exactly the same for all my other “rental” needs, such as car hire and broadband.
When I look at the cloud computing market, I want the same again. I want 10 Linux servers with 3vCPUs, 6GB RAM from a provider who guarantees PCI compliance and so on. Who can provide this and who is the cheapest?
Shouldn’t I be able to compare cloud service providers, as I do my energy providers? Shouldn’t I be able to switch between two providers quickly, seamlessly and without excessive personal involvement? This is the very near future of cloud computing services.
The next step – seamless migration between Cloud providers
Allow me to draw another analogy to the energy market. At the risk of banging my cost conscious mantra too hard, I must admit to being a member of an online energy club. This energy club is a switching service, which compares the market to show me the best deals.
It also goes one step further by sending me an email notification when there is a new provider that I could switch to and save even more money. This feature of the switching service allows me to set a threshold, say £50 per year. I receive an email when there is a new offer that would save me £50 or more per year and all I need to do is click “switch” to take advantage.
And the best thing? I also get cashback every time I switch. The “switching broker” provides me, the consumer, with a reward for switching providers. Compare this to a cloud broker and imagine a world whereby IT service providers pay you to bring your business onto their platform and the orchestration service makes this automated and painless; this world is fast approaching a reality.
Switching a single workload from one platform to another is fairly straight forward and there are some excellent tools that can help. But switching a database server from my 3 tier application stack to a cheaper provider in isolation is not a sensible move.
We need to consider data residency, move groups, network transfer costs and many other design principles before we migrate our more complex systems. Having said that, cloud brokering and orchestration services will soon mature to facilitate more complex migrations.
Increasing transparency & freedom in the Cloud Computing landscape
We need to consider the future impact of this. Computing is no longer a monopoly run by huge IT conglomerates, it is an open-book free market, whereby competition is maximised and the consumer holds more power than ever before.
Price comparison leads to price reductions, to price wars and ultimately IaaS will be provided on a price point which is equivalent to, if not less than, the cost price. This is not a bad thing, cheap computing with no capex supports innovation, powers start-ups and will lead the way to a new economy of interconnected, technology driven digital businesses and services.
Some service providers are already embracing this multi-vendor free-market that is developing for cloud services. Fujitsu for example, is committed to minimising vendor lock-in to solutions and already provides a built in capability to compare the top cloud service providers against a business requirement.
The output provides a graphical representation which can inform business users directly, or can be coupled with a cloud brokering advisory service to ensure that the best solution is selected, which meets the tactical and strategic aims of the business.
So what can we expect in the next 3 years?
- Cheaper pricing for IaaS, which reduces beyond the cost point in order to drive more usage of the wider platform capabilities in PaaS and orchestration services.
- Clients who actively avoid vendor lock-in and will shop around for the best deal, rather than remaining loyal to their current service partner.
- Automated migration services, which use algorithm based logic to prompt users to switch cloud services when business rules and thresholds are triggered.
- Automated orchestration tools, defined by business policies to minimise expenditure during non-peak hours. Automatically turning off workloads when the users go home at night.
- Service providers and brokers who pay the consumer to switch to their platform and ecosystem.