Australia’s data centre explosion continues at a pace not seen since the internet bubble, with new facilities coming on line each week. Australia is currently readying about 25 MW of data centre capacity, ranging from office block conversions to mega facilities. But is a data centre a commodity purchase, or something strategic?
Commodity: Pricing data shows considerable differences in the normalised cost of co-location hosting — nearly 300 per cent at the extreme — but the different charging regimes make comparison difficult. It is not a deliberate policy by vendors, but we advise clients to normalise the full lifetime costs to avoid cost penalties. IT costs are increasingly weighted towards OPEX rather than CAPEX. As power costs rise and carbon taxes remain a possibility, IT facility TCO becomes a key part of the decision. In some instances, it may never be possible to host IT within client facilities more cost effectively than in a co-lo; but don’t expect co-lo prices to plummet with increased competition and available capacity. Savvy operators will only bring capacity online just ahead of demand.
Strategic: Is a co-location decision like any other corporate real estate decision? The answer is “maybe”. In some organisations, IT growth can be highly volatile, liquid and fungible. In the financial services, for example, IT engines are cranked up when new products come online or new markets are traded. Clients in this sector see significant short-term system growth and rapid redeployment and, in this case, aligning the liquidity and volatility of the real estate to the IT silos removes the shackles of location constraints.
For non-dynamic organisations (if they still exist) IT real estate can be a simple binary decision. For many large organisations, blending owned and co-location capacity can provide long-term stability and short-term flexibility.
Cloud: As a data centre consultancy, 2010’s most common client theme was: “Does Cloud computing mean I won’t need a data centre in the future?” It is akin to me asking my wife, “If I have a garage, does that mean I don’t need a wardrobe?”
She would dearly like to see many items banished from my wardrobe; those trousers that now demand diaphragm contortions to wear, or the oh-so-bright patterned silk shirt bought on impulse in Asia. I’m fine moving those things but the English brogues and Italian silk ties stay where I can see them — and I will buy more in the future. I do need a wardrobe.
Steps in the data centre decision
Stop! Before you make your next data centre decision, we recommend the following five steps:
- Find out what real estate you actually have (most organisations are surprised at what they have) and compare it against what you need in the future. Consider growth, technology trends, business direction, regulatory environment, market competition and validate those assumptions with an independent third party.
- Obtain a clear understanding of how geographic dispersal affects your organisation for technology, operational, business and regulatory issues for all your data centre assets.
- Qualify your options. Understand what the market is offering for the price and confirm you will receive what you pay for. A five- or seven-year deal is a heavy financial burden without due diligence.
- Make the build versus buy decision within the financial, political and operational boundaries your organisation can realistically manage and always validate the decision with an independent party.
- Build the investment case with the income generating side of the business — it’s their money.
Simon Wise is a principal with CS Technology (Australia) a subsidiary of CS Technology Inc of New York. He has worked for CST for seven years in Europe, Asia and Australia with engagements across a broad range of industries including banking, government and education on data centre and technology projects.