""They 'airdropped' me into the middle of an ongoing discussion between the hospital and a vendor regarding the sub-par performance of a 3D scanner used to detect breast cancer," he told On Call. "When doctors viewed scans, they would sometimes appear in two seconds, and sometimes appear in 45 minutes," Warren wrote. When times ballooned to the larger figure, it caused all sorts of nasty scheduling problems as appointments ran past their allotted time."
"Warren started his attempt to sort this out by researching the network connecting the scanner, the storage device in one of the hospital's datacenters, and the medical suites where doctors tried to view the scans. He found robust 250 Mbps links and a two-hop connection across a small city. He then joined a conference call to discuss the matter and found it depressingly familiar. "The storage vendor blamed the client's network, and the client blamed the vendor's equipment," he told On Call."
A network engineer tasked with clearing a backlog investigated a 3D breast-scanner whose image retrieval times varied from two seconds to 45 minutes, causing scheduling overruns. Network inspection revealed robust 250 Mbps links and only two hops between the scanner, the datacenter storage, and medical suites. Vendor and hospital each blamed the other's equipment, but device logs showed no network errors. The engineer requested that staff report each occurrence of delay immediately. That reporting approach produced an incoming call the next day, suggesting user-observed timing events would be key to reproducing and diagnosing the issue.
Read at Theregister
Unable to calculate read time
Collection
[
|
...
]