PostgreSQL, Sea Lion, HTTPS, Authenteq, Tictail, Kademlia, Let's Encrypt, Sonera IPv6, Auth/Enc

Post date: Dec 18, 2015 4:44:07 PM

Excellent post about PostgreSQL Locking Revealed - Skip locked is awesome feature in PostgreSQL - Nice post summing up the most common lock stuff and performance issues like lock contention. I guess we've all had our moments with stale locks, race conditions and other not so amusing issues as well as with the additional complexity to deal with those.
I've been reading just so much project documentation that it's aww... Tiring. But there's no other way to catch up things, just read and think. That's the way. I don't know any shortcuts for that.
Watched Slush 2015 - slush.org - Watched opening and many pitching videos and grand finale of pitching contest.
Watched Jarno Limnéll's talk about Sea Lion submarine sea cable and it's meaning to Finnish Information Security. Personally I think it's important that we have multiple alternate paths to route the traffic via in case there are problems with some path(s). (C-Lion, Cinia) - Updated comment, this post is from my backlog. But right now, the cable is already being pulled from Rostock, Germany to the Île-de-Bréhat ship, as well as the splitting / branching / tapping point for Baltic states has been laid to the bottom of the sea. Sea Lion Log Book by Cinia (In Finnish).
Tons of fine tuning for libraries used for RESTful / JSON / SMTP / FTPS / FTP data deliveries with different integration cases.
Read long article about different means of hiding data in TCP, HTTPS (SSL), UDP packet headers, timestamps, etc. The places where most of people wouldn't look for it. Yet spotting that data is of course pretty easy for people whom are specialized spotting such commonly known channels.
Checked out Authenteq - Yet their website really lacks all details. I would like to see APIs as well as technical design white papers. It's easy to claim something without actual facts.
It's like the case of e-receipts. It's really trivial to deliver the receipts. As soon as you can easily identify the customer. That's the problem. As far I haven't seen any great solution to this problem. There are plenty of ways to do it, but all of those suck little or even more. Any of the numerous business partners I've talked about this topic, haven't been able to provide any good question. I've also received multiple queries from businesses working in this field if I would know a solution. If there would be a great solution, it should be patented or protected at least by NDA. But I can just say, there isn't one as far as anyone I know knows or tells. - NFC, RFID, Apps, Mobile, SIM, App, Loyalty Card, Barcode, entering email address, entering customer id / identifier / credentials or whatever.
Just watched future on E-commerce. Great talk. OpenBazaar fits great in that vision. They were mostly talking about Tictail which allows users to follow markets, etc. Now strong social network effects work in market places etc.
Studied Kademlia and reminded my self about it. All the basics have been clear for ages. More technical details. Also wrote some related code. I just love Kademlia, it's so beautifully constructed and considered. Especially the replication and caching parts are beautiful. Accelerated lookups is just basic binary tree stuff. I've also earlier checked out Chord, Tapestry and Pastry kw: Tonika, 5ttt
Let's Encrypt is entering public beta on 3rd of December 2015. Gotta try it out. Here's their blog post.
Here's a short document in Finnish about Sonera IPv6 experiences. - No news there at all. But good read if you're lightly technically curious.
With some projects I always get annoyed when people say something like authentication / signing being hard... It isn't. Challenge response with shared key works very well or signed data.Especially when considering not so advanced attacks, it isn't hard at all. It can be complex if the use case gets complex. But authenticating and encrypting data between two trusted parties isn't hard at all.