prohor

GCP also has sustained-usage discounts, which are very convenient, as even without reservation as in AWS, you get up to 30% discount if you have more constant usage, but you still have the full flexibility of on-demand.

When it comes to pricing comparisons of cloud see Cloudorado: https://www.cloudorado.com/

I have the same feeling, that the note notation is way too complex. I would love to see something like this getting traction.

Thanks. Yes, this is something what I think of. Does it show on the screen what operation you confirm with U2F? Or just "U2F authentication"? But cord makes it far less convenient unfortunately.

What I miss in the U2F tokens is a small display that would show transaction details which is approved.

Imagine you have a malware on your PC; it can send another transaction than the one you see in the browser, while the URL would still match. Having transaction summary on the token would be the last verification point where can still spot something is wrong.

The release is also improving sandboxing for Linux:

https://www.bleepingcomputer.com/news/security/firefox-57-br...

Sandboxing for Windows was introduced in version 54.

Rally nice analysis. It is a pity that open access to data like this is still exception, not a rule. I wish to have such data for my city.

Agreed - just check cloud comparison, AWS is rarely at the top: https://www.cloudorado.com/cloud_server_comparison.jsp

For me the difference seems to be in handling large number of distinct logs. In Kafka every log & partition is a separate file and moreover it keeps it open. So, storing multiple logs results in writing to many files so eventually random write IO; and also you may hit limits of open files. You can multiplex logical logs in each Kafka log, but then you read unnecessarily other logs.

Keeping SS tables makes it more sequential write and reasonably sequential write, as long as you have enough RAM to get multiple records of each log, so they constitute a continues blocks in flashed file.

Actually you could get very similar result using Cassandra, which also uses SS tables. The difference is that Cassandra keeps merging files, which actually makes much more IO traffic than clients. Cassandra will typically need 16x more IO for merging then actual data write rate. You can limit it a bit if you create time shard tables.

6 years ago I submitted cloud computing comparison - https://www.cloudorado.com/ . It has hit first page with 38 comments. There was a nice spike in traffic that I've never seen later but it faded quickly. Now traffic mostly comes from search and some links that popped up here and there. The site is live and provides revenue (but not spectacular; fraction of what I need for living).

There is similar plugin for Firefox for years. There is also one for Chrome that uses the same algorithm, so generates the same password. It keeps user name and version of the password for a domain in regular FF password storage.

https://addons.mozilla.org/pl/firefox/addon/password-hasher/

I just wonder - if the same key is used for enabling password manager and 2FA ... is it still 2FA? I mean, having the token you get both access to password and second factor to a service.

I love the feature they keep passwords in fact on the device itself, not as a key to enable password manager. I was looking for something like that. If only they offered strong encryption for Europe!

1.7 GW power. Wow! That means about 1 ton of CO2 released per hour (according to: https://carbonfund.org/how-we-calculate/).

The fun begins when you need to find performance issues in production, where you cannot really use profiler. Then you need to jump into APM tools. Unfortunately it seems there nothing really for that level for free, but take this for example: https://www.dynatrace.com/blog/code-level-visibility-for-nod...

In Poland there were many people killed, potentially due to a coma (,). A court process of the pacification of Wujek [1] was mainly whether a response to question "Should we shoot?" was "No, wait for orders" or "Don't wait for orders" (in Polish less difference "Nie, czekajcie na razkaz" vs. "Nie czekajcie na rozkaz"). This was really a ruining life of many just by one coma.

1. https://en.wikipedia.org/wiki/Pacification_of_Wujek

Thanks for hints. The Caddy server looks nice.

The default and simplest scenarios is that our domain is used, so that the user is not forced to setup DNS, but they can if they wish. But of course having a set of domains is an option. The problem with that is, that there is still a limited set of domains that we could use and still easily matches with the product.

It is about 1.5 month now.

Yes, using different domains is a potential option too, but the domain is directly connected with the product, so want to consider other options too.

Well, ACME would be perfect, but actually any fully automated process would do.

We provide an on-prem software available via browser. And well, we want to be very nice for our customers, so upon installation we also setup a subdomain in a domain that we control and request a certificate for that. At the end of installation we provide user with HTTPS URL where the service is available and with a valid certificate :-) Of course they can later opt-out, use their domain or certificate, but we make it work without security warnings from the first moment.

Spotify sounds very interesting to me. But I don't work there, so cannot really tell if it really works as they describe.

https://labs.spotify.com/2014/03/27/spotify-engineering-cult... https://labs.spotify.com/2014/09/20/spotify-engineering-cult...

I was always wondering rather on Stack Overflow for law, where one could get advice from other knowledgeable. But I talked with few lawyers (technologically open-minded) and they weren't interested. It seems the final root cause is that in software development the level that you get help is not the level where the end products are and compete. For lawyers it would be different - an advice is the root of their service, so it is the end product. So if they were helping with advice, they would directly help competition. Secondly, their customers could end up in that level instead (while software customers cannot benefit from stack overflow).

"large" is the smallest instance with fixed CPU allocation that AWS is providing these days.

I was quite wrong - IPv6 was just announced by AWS :-)