Laxintl From United States of America, joined May 2000, 25442 posts, RR: 49
Reply 1, posted (8 years 9 months 10 hours ago) and read 2062 times:
Odds are if you have a power outage other things at the airport such as the security checkpoint, jetways don't work either so its a mute point.
In case of particular carrier IT system outages I know some carriers have preprinted name list for flights, however most carriers will not attempt to operate with their IT systems down as things like printing bag tags, ticket issuance, passenger boarding, bagroom bag scans often cannot be completed without a working IT interface
Its a credit to the overall reliability and duplication of IT system's that its very rare significant disruptions ever occur.
From the desert to the sea, to all of Southern California
TWAAF9 From United States of America, joined Nov 2004, 88 posts, RR: 0
Reply 2, posted (8 years 9 months 9 hours ago) and read 2026 times:
Most airports have backup generators and can continue to function. The real issue comes when there's a reservation system outage. Manual check-in can be accomplished, but in order to comply with TSA regulations, all check-in must be completed remotely (at a location with working reservation system) to process all names through the No-Fly and Selectee lists.
Usually, the entire flight is checked in and a complete list of names is faxed to the station with the outage and the ticket counter agents hand-write boarding passes as each passenger presents him/herself at the counter. The faxed list also includes seat assignments, who has e-tickets, paper tickets, passenger final destination, etc. As check-in progresses, the agents simply check off who's showed up at the ticket counter. To ensure positive passenger bag match, any manual bag tags are also noted by each passenger's name.
At check-in cutoff time, the master list showing who's "checked in" and bag tag numbers goes to the gate. As each passenger boards the aircraft, the names are crossed out as being on board. In the end, you want the number of named checked off as "checked in" to match the number of names crossed off as "boarded."
After departure, the outage station then calls the remote location to reconcile who never "checked in" at the airport and the remote location returns the seat assignments. Then, if the reservation system still isn't functioning at the outage station, the outage station and remote location reconcile who actually "checked in" and "boarded" the aircraft: the remote location shows these passengers as "onboard" in the reservation system and closes the flight, thus processing the e-tickets.
ContnlEliteCMH From United States of America, joined Mar 2005, 1460 posts, RR: 44
Reply 3, posted (8 years 9 months 8 hours ago) and read 2005 times:
Quoting Laxintl (Reply 1): Its a credit to the overall reliability and duplication of IT system's that its very rare significant disruptions ever occur.
Indeed. Information technology has matured to a point of possible guaranteed continuous uptime; it continues to mature such that the cost of guaranteed uptime continues to decrease.
If you compare this to the aircraft industry, it's positively shocking. If an airliner has a dispatch reliability rating of 99%, that's acceptable, even though once out of 100 flights it doesn't go. If it does eight cycles a day, that means its out of the rotation once every 12 or 13 days for unplanned maintenance. It took the aircraft industry nearly 80 years to achieve this. This sort of reliability was simply impossible with reciprocating engines; it has only become possible with jets as that particular technology has matured.
Contrast this with IT. Assuming that the first computers were in use by the early 1970's, that means we're only 35 years in. But that's not a proper persepctive. Operational reliance on those computers, system-wide, is only about 20 years old. 1% downtime would bankrupt any airline. Why? There are 8,760 hours in a year. 1% downtime implies that it's okay to be down for 87.6 hours every year. If your math is poor, that's three and a half *days*. Acceptable downtime for these systems is probably 8 hours per year, or 1/10 of 1%.
An airline simply cannot operate without computers, which are required to sell tickets, accept revenue, schedule planes and parts, purchase fuel, board passengers, track bags, and pay employees. It's no exaggeration to say that an airlines computers are now as mission-critical as its aircraft! And if you don't believe it, ask Comair...
Christianity. Islam. Hinduism. Anthropogenic Global Warming. All are matters of faith!
Aloha73G From United States of America, joined Jul 2003, 2362 posts, RR: 4
Reply 4, posted (8 years 9 months 8 hours ago) and read 1995 times:
Last year when I was flying UA from HNL-LAX-SFO UA's mainframe in Denver crashed. It was about 6am at HNL with people checking for a 763 to SFO, a 763 to LAX and a 772 to NRT. All three flights were full. After waiting a bit to see if the computer would come back to life the agents separated the passengers by destination/flight and began manual check-in in the order of departure time (LAX, then SFO, then NRT.)
I was one of the last people in the LAX line and the computer came back to life right before I got to the counter. The agent was very nice and re-routed me onto the nonstop HNL-SFO It was interesting to see how they operated. The LAX flight boarded and left just a few minutes late and the SFO flight left ontime. The UA checkin staff in HNL did an awesome job!!
Aloha Airlines - The Spirit Moves Us. Gone but NEVER Forgotten. Aloha, A Hui Hou!
Bwaflyer From United Kingdom, joined Jun 2004, 689 posts, RR: 0
Reply 5, posted (8 years 8 months 4 weeks 1 day 23 hours ago) and read 1943 times:
In October I flew on SQ from JFK-FRA (v nice by the way) and they were operating a manual check in test. After weighing our bags on a manual scale, various people checked e ticket itineraries, and filled in a passenger record form which we kept with us until we reached check in. At check in, the agent called the SQ office (I'm assuming in NYC) to confirm my details and ticked me off against a manifest (which I assume has been faxed over to the airport earlier). Boarding passes were then hand written, and generic bag tags (stating either FRA or SIN) were attached to bags and the bag tag stuck to the boarding pass.
It took about an hour to check in, but the ground staff passed through the queue with drinks and snacks reguarly, letters with a full eplanation were handed out, and the station manager passed up and down the queue answering questions, and apologising for the delay.
Flashmeister From United States of America, joined Apr 2000, 2900 posts, RR: 6
Reply 6, posted (8 years 8 months 4 weeks 1 day 19 hours ago) and read 1884 times:
Speaking as an IT wonk, I can tell you that a number of companies (like airlines) strive for what's called five-nines for their mission-critical technologies. That means that these systems should be 99.999% available or reliable, depending on the measure that the company uses.
On average, this standard means that systems should not have unplanned downtime of more than 5 minutes, 15 seconds per year.
Obviously, maintenance, upgrades, and other work can bring a system down while the work is being performed. That's why the standard almost always mentions "unplanned downtime". The half hour that the entire thing is down to swap out a network switch doesn't count.
Companies spend millions upon millions to achieve five-nines, especially in industries like aviation, banking, and telecommunications. Airlines spend a virtual fortune on redundancy and resiliency, for good reason - no IP, no boarding passes, no departure, no revenue, and angry passengers. Even if the system may be slow, the system is at least up -- the standard generally doesn't say five-nines of great performance
Some achieve this standard, while others continue to throw money around. AT&T, for instance, had a standard of six-nines of availability when it built its long distance network. That's a failure rate of one in one million attempts. Their internet service, though, is lucky to see one-nine of reliability.