Cloud perspective to cloud services, CSA/SDP, BitCoin, NSA

Post date: Dec 21, 2013 6:52:37 AM

LinkedIn discussion about cloud services and my cloud perspective to it

I don't know if people in the group are going to appreciate this, but cloud is very different beast to different people. So I'm just going to stir the soup last time. Just to give bit more perspective in this group.

First of all thank you all, for this great discussion. Because it has been very interesting to read this thread. The viewpoint in this thread is totally different than in some other discussions. In those threds there are the actual tech guys talking, who are running the smaller cloud services talk about technology. Some smaller cloud companies are even run by a few guys taking care of one small data center, with something like 12 racks. It just still proves that cloud is something very versatile. They buy 1 unit servers, install Linux, use KVM or LXC with open source web mangement console to allow customer self-administration of used resources. As well as they also sell server administration to for the rented servers and for anyone else in need for managed servers.

Of course in those cases, security issues might be totally on different scale, when compared to major service providers. But that's just nature of the cloud, unless it's defined in more detail, it can be practically anything as I mentioned earlier.

With current open source software, it takes less than one week to get new cloud service up'n'running.

I'm quit sure many guys in this group haven't ever visited site called LEB, but it might be just nice refreshment for a busy day.

It seems that many people who are worried about the cloud, are talking about "enterprise cloud". Of course that's also different story.

Other stuff

• CSA - Software Defined Perimeter Finished reading
• Made one security audit in one case where I some how got feeling that everything isn't right. The end result was that 28% of 88 database servers were using default login & password. - I just loved it. This really can't be true, right? - But it is. So in theory systems are secure and security is tight. But the reality is just something very different.
• NSA using advertising cookies to track users? Hardly any news, I would also used what ever information is out there to identify users. Cookie tracking is nothing new, and if you're snooping traffic passing by, tracking cookies generated by other sites, is simply trivial of course. I've been saying that any site, which uses cookies or requires username / password or any kind of login, should always use HTTPS connections. Do people get it, nope. And this is exactly what follows from it.
• A nicely written article about how YIFY torrets is working around ISP censorship @ TorrentFreak.
• Why I didn't like BitCoin as design? I think it doesn't scale well. Because every user and node needs to know full history of all Bitcoin transactions. If Bitcoin is widely used, this quickly becomes really infeasible. And bit later confirmation to my thoughts from the Economist quote.
• Go users can now configure parallelism and concurrency option when using Go, they can set single App Engine instance to handle up to 500 concurrent requests, instead of the default value of 10 by setting max_concurrent_requests option.
• Finished reading a batch of New Scientist and The Economist magazines. I still got batch of system integrator and Service Oriented Architecture magazines to read.
• Quote from The Economist's BtCoin articld: "Bitcoin’s growing popularity is having other ripple e ects. Every participant in the system must keep a copy of the block chain, which now exceeds 11 gigabytes in size and continues to grow steadily. This alone deters casual use." - This is exactly the reason why I didn't like concept of BitCoin back in 2009 when I tried it. It was immediately clear, that it would lead to situation where number of BitCoin nodes is small and even if they claim that everyone can have a private wallet, it leads to situation where it's not true. If you're using BitCoin on your mobile device every now and then, it's non-feasible to maintain your own wallet. I've been reading and studying scalability a lot, and I don't like schemes which I can immediately recognize as non-scalable. Exactly same reasons why I didn't like original Bitmessage implementation & specification.
• Something completely different, just to have some perspective to other things: SR-72 project @ The Register @ Lockheed Martin @ Wikipedia,