Private cloud is the Reuben sandwich of clouds, or, Shafer’s Theory of (Private) Cloud
Software Defined Talk LLC·2017-07-14·via Software Defined Talk
Episode 99
·
July 14th, 2017
·
1 hr 28 secs
Microsoft will ship it’s private cloud stack, Azure Stack, in September. Will this work? Will people buy it? What could you even put in that cloud? You can feel that pull people have towards private cloud, so we’re looking forward to what happens. On a related topic, by our reckoning, kubernetes to small to have already fallen. Also: the elusive Baltimore accent, Oracle and containers, and recommendations.
Meta, follow-up, etc.
Where does Matt Ray find all these stories?
Patreon for this thing - like anyone who starts these things, I have no idea WTF it is, if it’s a good idea, or if I should be ashamed. Need some product/market fit.
This episode is sponsored by Casper, who’s looking for some good senior SREs. If you’re into building out and managing infrastructure that keeps code running and makes sure you can sleep soundly at night, check out the job listing, apply, and be sure to mention that you heard about it on Software Defined Talk. According to Glassdoor reviews, it’s a damn fine place to work. You can also just email jobs@casper.com and browse all their openings at casper.com/jobs.
Also, get $50 off mattresses with the code: horraymattray
Charges by consumption - how MIPS-y! “Compute charges start at .8 cents per virtual CPU per hour and go up from there, while storage starts at .6 cents per GB per hour. Those charges will be included in customers’ invoices for their overall use of Microsoft’s public cloud platform.”
Hardware via partners: “The exact pricing for Azure Stack hardware, including support contracts, will be up to each individual manufacturer. Microsoft is working with Dell EMC, Lenovo, HPE, Cisco, and Huawei to make the hardware available, and the first machines should be available in September.”
Scott Guthrie: “We talked to lots of customers who said, please don’t do that [allow so much customization that it's hard to debug problems]. The model we came up with instead was to work with a large spectrum of hardware providers, HP, Dell, Lenovo, Cisco and Huawei. Those are the five largest server manufacturers in the world. They will have systems that start with three nodes, not massive big purchases, that you can unbox and plug in. And have a fully working cloud in a day or two. Regardless of whom you call, we own the whole solution.”
Funny how OpenStack is now a cautionary tale. See Coté’s weasly, non-position on OpenStack in his May Register column.
I thought Cote’ was going to write this one up? (see below)
“Whitepaper review” a la The Weeds!
You have to be careful how you read that 451 survey.
On Asay’s 71%, Coté wrote: On that note, it’s easy to misread the widely quoted finding of “[n]early three-quarters (71 percent) of respondents indicated they are using Kubernetes” as meaning only Kubernetes. Actually, people are using many of them at once. The report clarifies this: “The fact that almost 75% of organizations reported using Kubernetes while the same group also reported significant use of other container management and orchestration software is evidence of a mixed market.”
Read: they’re trying everything. Nothing has won yet. Proving Asay’s point, but also defanging his link-bait lead.
“It seem far-fetched that Kubernetes could be heading for a fall” - there is no fall to be had because ascension hasn’t yet begin.
The core base of 201 people are organizations already using containers, so it doesn’t include organizations not using containers.
In a broader survey (where, presumably, not every enterprise was already using containers), of 300+ enterprises, production container use was: 19% in initial production, 8% were in broad production implementation.
This isn’t to say there hasn’t been huge growth in this space, but it’s the huge growth of small numbers.
This survey (though sponsored by CoreOS - I’m always suspicious of sponsored surveys, having worked on them myself!) is definitely worth paying attention to (as well as ongoing 451 and Gartner work here). Just make sure you read it right and don’t get too excited.
(Related: I’ve been thinking we should do special, “paid members only” [in Patreon?] “whitepaper review” episodes. Because, let’s be honest: only people who liked us enough to pay would be interested in that.)
Alright, now some vendor-sports:
So, can a vendor be successful if they “chase” the standards? Do you need to be in OpenWhisk, and OCI shit to operate in this space? Do you need to be Java EE compliant?
As John Willis would say, “I wanna be Ashlee Vance when I grow up.” Look at that guy: he’s kickin’ it no undershirt with the button-up style. Writes for Bloomberg, and only let’s just enough sass through in his tone to keep his broad, concise appeal but still have style that harkens back to his Register days.
And he wrote that Elon Musk book. CASH MONEY!
Second: “serverless is someone else’s server” manifest in 3D
Third, a lot of people ask me, “Coté, how do I get a job like yours?” Writing a lot of posts like this is one answer.
“The distribution users are, for most of the biggest projects, sysadmins.”
Nice history of how developers and distros have long been at odds.
Coté: I recall reading this. Did I get the summary right?
gems show that people will subvert the gate-keeper of being in the distro, thus, there is no power of a distro: people will just assemble whatever they want. Docker x10’s this.
Linux distro people fight too much and are, sort of, bags of dicks that are making the evolution of distros slow and further irrelevant. Hey, guys, “can’t we all just get along?”
Coté: thigh and leg chicken, BONELESS!, at CostCo. Enterprise Architecture as Strategy book - old (2006), but really, the topic doesn’t seemed to have change much, despite DevOps &shit.
