Docker Announces Commercial Partnership with Microsoft | Hacker News Reader

Docker Announces Commercial Partnership with Microsoft

blog.docker.com

44 pointshuslage13 comments

> With industry analysts declaring Windows Server with more than 60% of the x86 server market

Really? Where are such numbers published?

Yeah, this seems completely made up.

How come?

Actually this may be true, however... what they miss is, that these are mostly enterprise active directory servers which will never run docker.

> enterprise active directory servers

That's a small portion of all x86 servers. Statement like "more than 60% of the x86 server market" implies 60% of them all, which sounds very far off.

Well say it like that. Enterprises have probably a 80% Windows 20% Linux server, which results in something like 30%-40% of the total market + 20% servers that are serving the web / databases in the web whatever, which would result in 60%. And you can guess how many of these 60% would need docker. Probably less than 20%.

Its not surprising (except, perhaps, surprisingly low given the prevalence of free competitors) if, as is often the case, market share of X is calculated by ($ spent on X) / ($ spent on X and all competitors to X), and they are measuring x86 server OS license as the relevant market.

> if, as is often the case, market share of X is calculated by ($ spent on X) / ($ spent on X and all competitors to X)

That's a rather poor measurement method.

> That's a rather poor measurement method.

Its a reasonable and often (though not in this case, I would agree) the most interesting definition of share-of-the-market represented by a particular player.

I disagree. Such a measurement method has several significant flaws (which work to the advantage of whoever is doing the measurement):

1. How do you decide which price to count? Do you count initial costs, or all recurring fees? Do you count additional services by third party vendors? Do you count training materials and so on?

2. It ignores products that have no price, and emphasises inflated pricing of products. This is obviously massively incorrect as a measurement when it comes to services that can be done by gratis software (which includes a large amount of free software).

I think it's a relevant and useful measurement from the perspective of, imagine how much better free software would be if the cost spent per user were e.g. 1/2 (arbitrary) that of the Windows ecosystem cost spent per user. No doubt it's a lot less for free software, and that's fine. But that cost should be higher via companies and governments in particular, contributing more dollars voluntary, with donations and development for free software. Not merely free no cost, but libre free, as in repurposable.

Docker seems to be in great need for a monetizing play. Docker has a great mindshare in open source community but is making very less to none strides in commercial offering (read reaching profitability and keeping investors happy)...thus this PR!

Reply on news.ycombinator.com