Business Tips

Starting a business is 10% inspiration, 10% aspiration, and 80% perspiration. The first two, the motivational elements, are essential and need to be held within bounds, too little is a problem, too much disastrous.

Clearly there is no one-size-fits-all for business start-ups. A little thought throws up at least three possible scenarios:

1. The natural entrepreneur, the Richard Bransons, Henry Fords of this world, whose every bone is written through with the word 'entrepreneur' and has been since birth;

2. The professionals or tradesmen, the accountants, plumbers and builders who are naturally self-employed and for whom setting up a new business is a simple 'right of passage';

3. The middle-aged men or women who find themselves suddenly unemployable.

Each group is exposed to different risks. The first may go 'gung ho' at the whole business start-up, the second may suffer under the delusion that his basic skill - be it an accountant or a builder - equips him or her to be an entrepreneur, whilst the reluctant entrepreneur may be everlastingly looking over his shoulder for a more comfortable outcome.

The new entrepreneur must be a whole-hearted enthusiast for the project, but must serve up that enthusiasm with a significant dose of realism.

In every case the 'perspiration' has to be there from the start: save your 'castles in the sky' until after bed-time, every daylight hour has to be focused on hard reality. That doesn't mean that your start-up can not be fun: indeed it should be enjoyable. You will be confronted with issues you have probably never seriously considered before.

• How deep is your market? Will it support you and your family for the foreseeable future?

• How much money will you need to invest? When will you be able to start paying it off?

• What administration will you need? How much will it cost? How will you recover this 'overhead'?

The questions are many and variable and the answer to most is itself a question.

And as each answer is approached we need to maintain the magic balance: 10% inspiration, 10% aspiration, and 80% perspiration. We need to maintain our motivation but inspiration and aspiration on their own will never cut the mustard. We need hard facts and rational analysis to succeed; and we need them now, not in a few months when we have invested time, money and reputation and can see no way to retreat onto safer ground.

If it sounds like a battle, and looks like a battle, take it from me it almost certainly is a battle.

But the rewards of victory...

generic types of business models

The campus HPC centres mainly differ in their assumptions about the business environment and the way they capture value, how they finance themselves. There is a spectrum between two extremes: fully subsidised and fully sustained. On this spectrum, there are four prototypical business models of university HPC.

centres: library, shareholder, cost centre, and industry collaboration model.

These four generic models are described by outlining the basic idea, assumptions about the business environment, the financial model, and advantages and disadvantages of each.

1. Library model

The basic logic of this business model is similar to the one of a library: in a library a user can borrow a book from a basic collection without charge (if not yet used) in HPC service provisioning a user can consume an HPC service from a basic set of services without charge (if there is enough capacity).

2. Shareholder model

The basic idea of the shareholder model is that a group of ‘‘shareholders’’ who need computing power pool their money to finance a large common cluster. The HPC centre’s management or an IT portfolio manager conducts the acquisition, maintenance and operation of the system, which, in this model, is mostly a cluster

that can be scaled according to the needs of its ‘‘shareholders’’.

3. Cost centre model

Under this model an HPC centre is fully self-sustained and runs like an industrial cost centre with the need to compensate for all costs, including capital depreciation of the machines, staff costs, power, and energy. A researcher is charged a fee for an HPC service to cover all the costs, and whenever a researcher gets a grant, the money that he has been calculated for the facility goes straight to the HPC business unit.

4. Industry collaboration model

This model is based on a long-term collaboration: downstream, with an original equipment manufacturer (OEM) or upstream, by granting industrial partners access to the HPC infrastructure.