Buying a car is expensive. You have the initial capital cost of buying the car and then the operating costs of petrol/diesel/electricity, insurance, maintenance, etc. as you use it. OK - no surprises on the tradeoffs around making such an investment decision. But what would your decision be if at some point within the next two years all the roads, garages, petrol stations, (i.e. the supporting infrastructure) that you use in your area were going to be removed? You might rethink your decision to buy a car.
A similar situation is facing researchers. In the past, in order to do research you need infrastructure - a roof over your head, access to a library, a laboratory, an office, a desk, etc. Increasingly, you need e-infrastructure - access to a computer, the internet, software, possibly a high performance computer or two, data sources, data storage, etc. We don't realise how much we have become to depend on such e-infrastructure until it goes away!
So, if you are using e-infrastructure as part of your work, you want to know that its going to be there and that its going to work! Research grants generally only have funding that lasts between 3 and 5 years - even if it is part of a longer research programme. Making a decision as to which e-infrastructure resources you are going to use and depend on for your research becomes a critical decision.
E-infrastructure can be seen to therefore have several components:
- the software that you use locally to access your e-infrastructure resources
- the software deployed on physical resources to expose them as e-infrastructure resources
The quality, reliability, scalability and sustainability of e-infrastructure are big issues for further discussion...