Security, security, security: What a week of outages means to IT (and you)

Computerworld ·

Security, security, security: What a week of outages means to IT (and you)

The last week of major security and resilience problems is proof that business and consumer users must think just as deeply about tech resilience as any big enterprise. The pace at which major outages have impacted major entities in the last few days will already have had an impact on business, imposing real consequences. A week IT leaders can’t ignore The incidents are very hard to ignore. Microsoft’s major outage November 12-13 took out Microsoft 365 services for an unacceptable 24 hours , and the FAA grounded US flights for more than four hours on Nov. 13 because of network equipment outages. This week, a JPMorgan Chase outage rippled throughout the financial system, United Airlines experienced a widespread system outage , the US National Security Agency’s classified internal network went down , and then many of Apple’s own systems also hit problems. (Officials do say all these incidents were not connected other than by temporal proximity.) Apple’s big outage saw several of its enterprise and consumer services fail. These included iCloud sign-in and backups, Apple Pay, the App Store, and Apple Support. It also extended to developer services such as App Store Connect and Xcode Cloud and the company’s own internal corporate tools, including AppleWeb. Internal system failures meant physical Apple Store retail operations were also hobbled, disrupting sales, customer pickups, and device repairs. There were reports that device management onboarding was briefly out of action, while others reported errors setting up new iPhones. Apple moved quite swiftly to resolve the problems inside working hours, from when they were first reported at around 11:45 a.m. ET to full restoration at around 6 p.m. ET. So, was it a hack? Apple said it was not caused by a cyberattack and has given no indication that personal data was put at risk. The company told NBC News that the disruption was caused by an internal update but didn’t say anything at all beyond that. “Many of our systems are back up and running after an internal update caused some outages this morning,” Apple said. “We’re working to get the remaining systems online as soon as possible.” Whatever Apple did seems to have worked. As of this morning, DownDetector shows Apple’s services working normally even though Apple’s developer services page shows an outage with App Store Connect. We don’t know what failed or why the update generated what could be multiple failures across Apple’s systems, each of which rely on their own infrastructure. We don’t know why the problem wasn’t caught in staging or contained within a staged rollout. Nor do we know if the failure took place at a shared layer, such as identity or authorization, and we aren’t likely to know unless Apple publishes a post-incident report. Shared layers, multiple dependencies That leaves us with conjecture. One possibility: the manner in which the failure manifested itself points to a shared dependency. Sign-in, payments, developer authentication, and internal staff tools all depend on authorization , so a problem there may have taken down services with otherwise unrelated back ends. A fix applied to a shared layer would fit a recovery by close of business, but so would a rollback of whatever update Apple applied. At the same time, the cascading nature of the problem shows just how much we all rely on digital infrastructure and the extent to which any failure in those systems threatens businesses and end users. The timing of these incidents couldn’t be worse as global governments continue to tell citizens to prepare bug out bags for use in emergencies, raising the general sense of fear and uncertainty. Resilience isn’t glamorous Apple hasn’t said whether it will publish a post-incident report, so unless it does, the safe conclusion is that an internal change broke something shared, that personal data appears unaffected, and that the rest is unknown. But what isn’t resolved is how consumer and business users are meant to protect themselves and their own business continuity when two of the world’s leading computing platform providers hit serious outages within a week. We must get better prepared for such challenges . The practical response is to identify which critical processes depend on a single provider, maintain a recovery path that doesn’t run through that provider, and be ready to distrust unexpected “fix it” messages . Staff locked out of sign-in or any other service should be encouraged not to use unsanctioned tools or share credentials but to wait for the situation to be resolved or use a sanctioned fallback if one has been specified. Now please subscribe to my daily, human-curated Apple-related news headline feed at The Core , or follow me on BlueSky , LinkedIn , or Mastodon .

Источник: Computerworld