By Roger Lawrence, CTO Stategic Enterprise Services - HP South Pacific
Elasticity is one of the most compelling promises of (public) cloud computing. Speak to any public vendor, or their (start-up) customers, and they’ll wax lyrical about the benefits of public cloud. They will speak of not having to pay for what you don’t need, scaling automatically on demand and scaling back when workloads change.
But does this also apply to the Enterprise?
More importantly, what is the business impact of elastic compute delivery?
I recently conducted a CIO Vision workshop with a client of ours. As part of the workshop we looked at this particular organisation’s IT spend as a percentage of revenue, with two comparative points:
- Historic spend vs. current spend—This shows whether a company has been investing/overspending or benefitting from efficiencies/under investing
- IT Spend ratio vs. competitors—Again this can show interesting insights
What was interesting is that this organisation spends a miniscule percentage of their revenue on IT. The company’s spend is down in the single percentage point range. Their biggest spending competitor had an IT spend of about two percent of revenue.
So even if we managed to save 10 percent of their IT spend across the whole business, this would be negligible and could be considered a rounding error.
Don’t get me wrong, I believe in saving operational expenses and driving efficiencies. I also believe we should use capital effectively as much as possible, and not simply invest in assets that we can’t use to drive revenues. Yes, it would be good to shift the greater part of the IT budget from business as usual (BAU[KJL1] ) or “Keeping the lights on” to innovation.
I am suggesting that an enterprise has a far different usage profile than a small or start-up business. But I still agree that the large enterprise can benefit from elasticity, at a business level.
Let’s look again at our workloads more closely.
There are some workloads that are core to the business. The inventory-control system for a manufacturing plant, or the logistics system for a trucking company. Without these systems, the business can’t operate. Or at least they can’t operate for very long.
By and large, these systems also don’t have peaky workloads. They scale predictably, if at all, and a business can realise the benefits of depreciating their capital investment over a number of years.
But even large businesses do have workloads that are seasonal.
An example of a seasonal enterprise workload is the trading systems for an investment bank. These systems need to cope with increased loads when companies make their Initial Public Offerings. Or, let’s say the marketing department runs seasonal campaigns or accounts needs to deal with quarter- and year-end-processing—even the Point of Sale system for a retailer can be seasonal.
If we can bring small company agility to the enterprise, we have an opportunity to make a substantial difference, not only to costs, but also to the revenue, profit and cash flow.
As part of the overall IT strategy, look at the various applications and systems that support seasonal, or really dynamic workloads. Consider how to deliver these applications through cloud computing architectures. Keep in mind that this looks different for different organisations and different demand loads.
Perhaps this is through a privately owned, standardised and automated infrastructure platform, e.g. HP’s Cloud System offerings. Perhaps it’s using a virtual private, or even a public cloud offering. Or maybe you own infrastructure for normal workloads and “burst” to various cloud providers as necessary.
Let’s look back at the client’s organisation that I mentioned earlier. It turns out that reducing their “Day Sales Outstanding” by just 1 day would improve their overall cash flow by more than the entire annual IT spend.
This is where elastic cloud platforms, driving agility, could really benefit the enterprise.