Blog‎ > ‎

Archive, Networking, CPUs, Snailmail, Kill Switch, Backups, Locking, eReceipts, UPS & Power

posted May 20, 2018, 12:54 AM by Sami Lehtinen   [ updated May 20, 2018, 12:55 AM ]
  • I've got a site example.domain and I've also got a www.example.domain subdomain configured. Only the example.domain is actively being used. Therefore the www.example.domain servers robots.txt with disallow: / . For some strange reason - web archive / wayback machine - crawler seems to think that this should be applied to example.domain which naturally doesn't have such a restrictive robots.txt. It also seems that web archive is the only crawler doing such a silly assumption. Why they're doing it? Isn't that a bug? It should be pretty obvious that it's not the same, and the robots.txt shouldn't be applied like that between two different domains - naked and www one.
  • Reminded my self about Cat7 (class f) cables and EuroDOCSIS 3.1 as well as 802.11ac technologies.
  • Multi-threading is coming more important everyday. Intel just announced desktop i9 CPUs with 20 threads. Also latest ARM processors can go up to 64 cores at least on specification level.
  • Checked some statistics, it's totally incredible that in Finland there are 34 million parcels processed in mail per year. And 10 million of those parcels are from China. That's truly incredible ratio. No wonder it takes sometimes a while for customs to process those. Especially under xmas etc.
  • WanaCrpyt0r / WannaCry discussion, domain checks, anti sandbox measures, 'kill switch fail' etc, bad security, internal networks. It's funny how many successes and fails can fit into 'one project'.
  • Microsoft Kill Switch. Yep, that's what I've been saying and thinking for a long time. People don't realize they're using backdoored and booby-trapped software. Which can be leveraged at a right moment, if required. Boom, and you're totally screwed. They can trigger it at anytime. It's not a joke that Windows is very expensive malware / virus. Quotes: "Expensive and hampering, Unfair competition, Intense lobbying, Security risk".
  • It's wonderful when document & SQL databases are mixed in sweet harmony. It's not hard, but it seems that for some parties it's easy to mess-up things. What's the difference of PUT, PATCH, delta UPDATE, UPSERT with absolute value, etc. Sigh. Not that hard after all, but it's important to get those things right. Otherwise end result will be ... interesting ... Yet if the testing is done in single thread mode, without nasty edge cases, it will probably pass and everything seems to work. Until race conditions and 'unexpected' errors start to occur. In this sense, even if you wouldn't personally use different kind of databases or solutions, it's very important to know how the alternative technologies work.
  • Good day of hard work, automated Duplicati 2 backups, backup integrity testing, and data verification. Phew. What's done is done.
  • Had a long discussion about physical security and locking with one organization. It's just as important with physical security that things are designed right, as it is with information systems security. If design is bad, and users get seriously annoyed by it. It's highly likely that the security features won't get used at all, because people prefer disabling such features. Even if those would make them more safe. Yet more safe for random rare events isn't great trade-off if it requires annoying daily tasks requiring time and effort. I guess this applies to most of security stuff anyway.
  • The Quick vs. the Strong: Commentary on Cory Doctorow's Walkaway - Nice essay by Bruce Schneier. Times change. New platforms can be a wild west for a while. But if those become popular. Yep, nothing new there.
  • One e-Receipt service I'm using, lags often for 15 - 30 minutes. I find that incredibly annoying. 15 - 30 seconds would be ok. But it's still a long time, if you don't know if you're going to get a receipt of the transaction or not.
  • Microsoft Office 365 / Microsoft account password change took once again more than 6 hours to be effective. Afaik, I personally find that being a serious security flaw, allowing access with old password for that long.
  • Run once again the regular UPS powered checks, that all key systems remain working for at least 24 hours after power is cut. Also alerts are automatically sent. But that gives us reasonable time to react even during weekends etc. Anyway, it's very rare that power outages last more than a few minutes when power is getting rerouted. Usually there isn't even one power outage per year. This is naturally limited to the vital systems, all secondary systems shutdown start shutting down after a few minutes and there's generator power for those. But that's totally acceptable risk in this kind of environment / use case.