I've worked in IT for 36 years, and worked on very large systems in companies like BP North America, Conoco, Currency Systems International, Buckhorn Petroleum, Exeter Petroleum, Quinoco Resources and others. I've not only done programming, but I've maintained accounting system, advised on hardware and installed hardware. One of the primary things that I've done is work on disaster recover systems - otherwise known as DR systems.
I can just about give you a 100% guarantee that there was no 'power outage;' only a virus attack would bring their entire system down. EVERY major company has a DR plan in place. DR plans, for major companies, require a backup server for every mission-critical server, and those servers are usually located in another city (if not another state). These servers do what's called a 'fail-over,' so that when the primary server fails, the DR server online automatically. Even if EVERY server's power, in every city, went out, the DR servers would allow operations to continue without so much as a burp. It would take all mission-critical servers and their associated DR servers going down (or being corrupted) to cause a system 'crash.'
For instance, when I worked at BP North America, in Houston, TX, we would hold DR drills every few months, and the DR team would fly out to our DR locations and BP would then take some or all of Houston's mission-critical servers offline - just like what would happen, for example, if a hurricane caused widespread power failures. The switch-overs were almost always seamless. Do you really think that Delta, who's servers are WAY more mission-critical, doesn't have at least as robust a DR system?
The DR servers, btw, are updated in virtually real-time, so they contain a 'mirror' of all of the information on the critical primary servers.
I am of the opinion, given 36 years in the IT business, that only a virus that had been in place, perhaps for months, would cause a system-wide crash.
If, however, the virus had a delayed action, the DR system would have also been compromised, as well as backups going back to the date of infection.
I can't begin to see how they would recover from that without SERIOUS economic impact. They could possibly lose every booking, every payment, every vendor order (and payment) and every employee transaction back to the date at which they were first infected. Additionally, the virus probably dumped, right before corrupting the system, all passenger an payment information. If the accounting system was also compromised, then all of their accounting information, such as back account info, vendor information, trade secrets, etc. would be available to identity thieves (for a price.)
If I were going to write the virus, I would get it installed and it would then self-replicate across the system - and hide. After 60 days, THEN the virus would go active. It would transmit any information that I've asked it to gather and send it to me, and then it would encrypt or wipe the hard drives.
The DR system would, thus, also be compromised, and you'd have to go back 60 days before you got clean programs and data. All reservations and, possibly, all accounting data, would be as if was never entered for the past 60 days.
Ok, so a company, also, will completely back up all programs and data every few days and send those backups off-site to a secure storage facility. That's good, right?
Any attempt to restore from off-site backups that contained the virus would trigger the virus as soon as the program went live and the virus recognized the current date.
So long, Delta!