Assets over certain value are depreciated over time for the company records. A public company needs to prove that these are not fraudulent assets. If you can't show the asset which you cannot find it, then you have to write it off. This becomes a true financial pit for an organisation.
Have the new trends in datacenter thus enhanced the role of CFOs?
Hsu: It depends on the size of the organization. For our client eBay the largest capital expense ever is across their datacenters. The expense line certainly rides to the top and CFOs certainly wear it. Very often with our clients, a business unit will need five racks or a row of cabinet which is a capital expense large enough for a review by C level -- not necessarily CFO level only. If you can delay that capex spend by one year then ROI can be extremely high. You could virtually pay for anything from technology point of view whether it is Software or hardware solutions or combination of both to delay the expense.
How does delay of capex to rebuilt or enhance a datacenter actually benefit the organizations?
Hsu: A datacenter has certain planned capacity usually measured in energy say 200 kilowatt or megawatt. It used to be in rack space but that changed a decade ago. However adequate the client does the planning at some point they know they have used all the capacity and they need to build more capacity. Universally the clients go wrong in almost each time.
If they have 1 MW and they need 200 kW more, in almost every case they don't actually need to build that 200kW. Building that extra capacity is extremely expensive. Any vendor like Emerson, Eaton, Raritan pursue that logical defense for the organizations to build extra capacity when they really do not need it. By delayed expense, what I really mean is to use the 1 MW that you already have and then you do not need to build more capacity now till you actually need it.
What pitfalls should CIOs avoid before executing the DCIM strategy in their organization?
Hsu: The first one is the reality around expectations management. There needs to be organisational buy in and the recognition that it can be done in a better and systematic way. You can't throw thousands of dollars to implement an optimised software to solve DCIM problem. You have to connect the actual owners of the business processes - the people who run the batteries so as to speak. They have to lean and understand to ensure a successful software implementation. Otherwise it often tends to be waste of money.
Sign up for CIO Asia eNewsletters.