Cut off in the seclusion of space, crew members living aboard the International Space Station (ISS) depend on regular deliveries of air, water, food and fuel for their survival.
But when the ageing space shuttle fleet is retired in 2010, the US space agency (Nasa) will lose a principal means of ferrying crew and cargo to the ISS.
The shuttle's replacement - Ares-Orion - will not enter service until 2015 at the earliest.
And in April, Nasa told legislators it would stop asking for Congressional permission to buy cargo space on Russian Progress re-supply vehicles after 2011.
That leaves the US dependent on European and Japanese spacecraft for delivering supplies to the space station.
But Nasa has also been pursuing a commercial approach.
Three years ago, the space agency took the unprecedented step of fostering the development of private spacecraft designed to carry crew and cargo to the ISS.
It offered $500m (£340m) in "seed money" to help stimulate a competitive market for supply flights to the space station.
This month, Nasa awarded two companies - SpaceX and Orbital Sciences Corporation - with lucrative contracts to provide cargo delivery flights to the ISS up to 2016.
Nasa's administrator Michael Griffin said he hoped the commercial ventures would succeed. But he told BBC News recently: "It's not commercial if Nasa is sitting around telling them what to do and how to do it. I don't think they need that."
Elon Musk, the South Africa-born entrepreneur who co-founded SpaceX, says: "Even when [Ares-Orion] does come online, it's sort of overkill to use it for servicing the space station. It would be incredibly expensive. So Nasa looked to the private sector to solve its problem."
Mr Musk made a fortune from the sale of his internet payment service PayPal to eBay and has invested at least $100m (£68m) of his own money in SpaceX, based in Hawthorne, California.
The company's cargo re-supply plans are based around a rocket called Falcon 9. The standard version of the Falcon 9 is arguably a medium-lift launcher, says Mr Musk, designed to place 9,900kg into low-Earth orbit (LEO).
Assembly of the first Falcon 9 at SpaceX's new launch site at Cape Canaveral, Florida, should be complete by 31 December 2008.
"To external appearances, it looks like something from the Apollo programme, or Gemini, or Soyuz. But the materials are much more advanced, it's designed to be reusable - which is an important characteristic," Mr Musk told BBC News.
Falcon 9 is the intended launch vehicle for a capsule, called Dragon, measuring some three and a half metres (12ft) in diameter. Dragon is designed to carry more than 2,500kg of cargo, or a crew of up to seven, into LEO.
It is capable of carrying both pressurised items - those that need to be kept at Earth pressure and are to be used inside the space station - and unpressurised cargo - to be used outside the ISS, such as control moment gyros.
"The Falcon 9-Dragon system is intended to replace the function of the space shuttle when that retires in 2010," says Elon Musk.
Falcon 9 will place Dragon in an initial parking orbit. From there it will manoeuvre towards the ISS. Dragon will make a slow approach and, once in range, will be grabbed by the space station's robotic arm and berthed.
During the high speed return to Earth, Dragon will be protected by a heat shield made of phenolic impregnated carbon ablator (PICA). This highly heat-resistant material is barely scathed at heat fluxes that would vapourise steel.
The capsule will parachute down to the sea for recovery.
Safety is of paramount consideration: the manned version will have an escape tower to rescue the crew if something goes wrong - a feature absent from the space shuttle."
"Hopefully we'll do the first demonstration flight next year of the Falcon 9-Dragon system, then particular demonstrations in 2010 and start doing operation missions possibly by the end of 2010," Mr Musk told BBC News.
First of all, Dragon will carry astronauts from Nasa and from other space agencies to the ISS. But Mr Musk hopes also to transport space tourists to private orbiting stations.
One company, Bigelow Aerospace, is planning to assemble an orbiting "space hotel" based on a series of inflatable modules.
"We have also thought of perhaps carrying private space adventurers on a loop around the Moon," says Mr Musk, adding that this would probably cost on the order of $40m-$50m per person.
"I think there is a wide range of applications. Perhaps the Falcon 9-Dragon system will ultimately evolve into something that will take people to Mars."
'Big empty can'
The other winning bid in Nasa's cargo re-supply contract was made by Orbital Sciences Corporation, based in Dulles, Virginia.
Orbital's vehicle consists of a medium-lift rocket called Taurus 2 which will be used to launch the Cygnus capsule. Unlike Dragon, Cygnus will only carry cargo - not astronauts.
Launching from the Wallops Flight Facility in Virginia, the medium-lift rocket will be able to ferry 5,500kg into LEO. Orbital is due to carry out a demonstration flight in the fourth quarter of 2010.
"We took a bunch of existing parts but combined them in a way that is very, very efficient," says Dr Antonio Elias, Orbital's executive vide-president, who is overseeing the development of the new system.
Cygnus is based around a common service module, containing the vehicle's propulsion, power systems and avionics. To this common module is added one of three types of specialised cargo modules - each designed for different mission scenarios.
One of these specialised modules will carry pressurised cargo, another will transfer unpressurised cargo, and a third type of module will return cargo items from the space station to Earth.
"The one that will be used the most, I believe, is the pressurised cargo module," Orbital's executive vice president told BBC News.
Dr Elias describes this module as a "big empty can". It is "volumetrically efficient and light" because, says Dr Elias, "all it has to do is bring cargo up".
"It gets attached to the station, the hatch opens and crew empty the pressurised 'can' of its contents. They fill it up with trash, close the hatch. The service module backs it out of position and de-orbits it over the Pacific Ocean. Both can and service module then perish in a fiery ball of plasma," Dr Elias told BBC News.
The unpressurised module is less efficient because some complex, heavy mechanisms are required to attach cargo: "The boxes you carry have to be very far apart because the (ISS robotic arm) has to come around and handle them. You have to give it lots of clearance," says Dr Elias.
The efficiency of the cargo return module, he says, is relatively low because of the shielding, parachutes and other paraphernalia required. He expected only a fraction of re-supply flights would require the use of this module.
Orbital says this approach of using specialised modules keeps development costs low and counter-acts some of the forces which drive payloads to be launched on heavy-lift rockets..
Dr Elias was chief designer of Orbital's Pegasus rocket, the first privately developed launch vehicle, which made its maiden flight in 1990.
A few years ago, he says, Orbital came up with an idea to re-supply the space station using Pegasus. But, at the time, the benefits were not clear to either Orbital or Nasa.
In fact, the origins of Orbital's Taurus 2-Cygnus system can be traced to the demise of the Delta 2 rocket.
For two decades, the Delta 2 had been the US fleet's most reliable medium-lift launcher for military, scientific and commercial payloads. It is still a perfectly good rocket, but Nasa plans to make its final Delta 2 launch at the end of the decade, shifting more of its medium-lift launch traffic to the Atlas 5 or Delta 4 heavy launch vehicles.
The US Air Force's Evolved Expendable Launch Vehicles (EELVs), though significantly bigger, are expected to be comparable in cost in the long run - if not cheaper - than the Delta 2.
"We became concerned that the US government satellites we were bidding for, winning and building in this class would disappear for lack of a launch vehicle," Dr Elias explains.
"We were concerned this would favour the larger spacecraft launches on [EELVs] and that the market would go to the big companies, such as Lockheed Martin, Boeing and Northrop Grumman."
Orbital began designing a successor to the Delta 2 with its own money. Although the company was on solid financial ground, finding a market to justify the expense was not easy. But a new opportunity was about to present itself.
In August 2006, Nasa selected two companies - SpaceX and Rocketplane Kistler - to develop and demonstrate orbital re-supply vehicles under its Commercial Orbital Transportation Services (COTS) programme, setting the firms aggressive timelines.
The space agency dropped Rocketplane Kistler a year into the programme; the company reportedly failed to meet a development milestone to Nasa's satisfaction.
"We decided to make an offering whereby Orbital would provide out of its own funds not only the additional money to develop Taurus 2, but also a space vehicle that would be suitable to provide those services," said Dr Elias.
Orbital filled the void left by the departure of Rocketplane Kistler, winning a Nasa contract under Phase II of COTS.
Each of the Commercial Resupply Services (CRS) contracts awarded to SpaceX and Orbital in December is worth a potential $3.1bn (£2.1bn). But the market itself remains very small.
"I don't think the market can support more than two companies. And it's going to be hard for it even to support two," Dr Elias told BBC News.
"However, as prudent businessmen, we did not embark on this venture believing we would grab 100% of the demand. So we are willing to be profitable in a situation where we only have half of it."
Observers point out that Nasa is betting on vehicles which do not yet exist, an approach which presents a major risk for the space agency.
Not only is it relying on two companies to keep supplies coming to the ISS, Nasa hopes the rocket and cargo vehicles can be developed in months - not the years it has usually taken other agency programmes.
"The task is significantly bigger than anything either company has ever done," John Pike, a space policy analyst for GlobalSecurity.org, told the LA Times.
"All of these things strike me as significant challenges for even the biggest aerospace companies."
But Nasa is not putting all its eggs in one basket. It can still barter for cargo space aboard the European Automated Transfer Vehicle (ATV) and Japan's H-2 Transfer Vehicle, or HTV, which is due to enter service in 2009.
Europe also plans to modify the ATV so that it can bring cargo back from the space station, a capability Nasa is eager to have.
Announcing the award of the CRS contract, Bill Gerstenmaier, Nasa's chief of space operations, said: "This is a pretty monumental thing for us, this is a contract that we really need to keep space station flying and to service space station."
He added: "I think it's exciting we're doing this from the commercial side. We've got some good proposals and we've chosen the two winners."