I was on the design team for the MiSeq DNA sequencer at Illumina that can sequence 1 billion bases in one day, doing embedded systems/FPGA/control loop work. I no longer work there, but think they've managed to increase throughput. This particular unit fits on a tabletop, and costs about $100K.
A story was related to me while working there about an outbreak in the intensive care unit in Cambridge England where 7 preemie infants got sick. With this instrument, they could see how the virus mutated on a room-by-room basis, and a day-by-day basis. It was apparently unprecedented. They had one of our instruments on an early trial basis to give feedback on it's usage. The pathology department was pretty excited. This seems like a very useful kind of instrument when tracking the spread of diseases. I'd be curious about the adoption rates for such instruments in pathology labs, the CDC, etc. I understand that Illumina has made a push to have their instruments certified as a medical device, but I don't know the status of it. I'd like our labs to have all the tools they need to rapidly converge on the infectious agent, etc.
One important consideration for portable DNA sequencers is the read error rate of the DNA fragments (akin to bit error rate in a length of magnetic tape). The higher bit bit error rate, the more samples you have to make to reduce the probability of error to a small acceptable level. Even though some instruments on the market may be cheaper to run, you have to read a lot more samples to reduce the error statistics. (the Q scores). Any portable instrument must do this with a low error rate, such that the small sample size is meaningful. Also, the longer the read length of an individual strand the better.
DNA sequencing is sort of like taking a photograph and cutting it up into thousands of pieces, and reassembling it. The bigger the chucks, the more distinctive it is, and the easier it is to fit into a larger puzzle, pieces that are too small, like bits of sky aren't distinctive enough to see how they fit into the larger picture . I still don't think we've been able to completely DNA sequence a human being, because the "sequencing-by-synthesis" method used by Illumina only uses relatively short strands of 100base pairs (more if you do "paired-end" sequencing that pushes it to +250, though my knowledge is a few years old).. There is some small percentage that they can't fit because it's not distinctive enough, and the DNA itself does not break apart uni-formally. Some areas are over represented, and other ares where they're underrepresented..