What kind of VPS models is it you (as a customer) want these days?

doghouch

@Brad said:
failure

I smell @goodhosting's alt

Ympker

Thank you guys again for all the input

Jonchun

@TinyTunnel_Tom said:
A chair in CC's NY is about as much use as a fucking server there.

Lower power costs. Roughly the same throughput... I see your point.

inthecloudblog

abusable cores and free incoming bw. no need for:

more than a single ipv4
for ipv6
ssd
superb uptime

Ympker

@doughmanes said:
Unique control panel/billing area + pooled resources and reseller enabled. It beats developing plans to fit into customer's budgets versus just selling them resources. Want 4 CPUs, 8GB of RAM and 1TB of disk so a customer can create 2 VMs? - sure $x.xx due. Oh 4 CPUs, 16GB of RAM, and only 250gb disk? This much.

Being a Webdev I am currently working already on a custom cp along with billing system Custom CP will have alot of features (which I am already adding on the side) such as: Start/Stop/Reinstall server, 1-Click installation of various webapps, stacks, gameservers, os templates. Then of course an in-panel terminal access, dns managenent, ftp browser, reseller panel (afflink creation), possibility to install various ovz/kvm instances on a dedicated with one click (assign ressources, os and ip to them) and assigning them to your customers, in-panel knowledgebase, ticket system (for your customers), display of various stats such as cpu usage, ram usage, bandwith usage, possibility to limit your customers to bandwith, possibility to upgrade your vps resources with a click of a button...

GM2015

You left me in suspension. Can't wait.

Ympker said: with a click of a button

Anyway, sarcasm off, good luck with it.

Ympker

@GM2015 said:
You left me in suspension. Can't wait.

Thanks

ceibaNet

the plans you're offering look to be pretty good. A nice variety to choose from. But also remember they should scale to maybe at least 2 more levels

Ympker

@ceibaNet said:
the plans you're offering look to be pretty good. A nice variety to choose from. But also remember they should scale to maybe at least 2 more levels

Thanks for the feedback. Yes, I already planned to have 2-3 more upgrades

kingcobs999

@Kris said:
If you're studying now, by the time you start hosting or offering it, it will all be about Docker, Kubernetes, containerization really.

The industry is waking up about benchmarks mattering when scaling, and Xen / even KVM just has too much overhead.

It's all about Docker / Kubernetes going into 2016, with more adoption coming in 2017 likely. With Google backing Kubernetes, I see that going far.

Are you James Bottomley by any chance? Because he's the only one who makes such claims about containers on every chance he gets.

Kris

No, but one day after I said that, looks like Docker purchased Unikernel.

So why is Docker interested in all of this? Docker founder and CTO Solomon Hykes acknowledged that this is likely the “most obscure” of Docker’s acquisitions, but he also told me that he sees it as the company’s most exciting one to date.

The 13-people strong Unikernel Systems team is largely comprises developers who previously worked on the Xen hypervisor. Unikernel Systems is a major contributor to the overall unikernel ecosystem and its open source components. Hykes tells me Docker will continue to be very active in this community.

With this acquisition, Docker is bringing a lot of deep technical knowledge into the fold. “Expect the Docker platform to be much more aggressive in solving problems lower in the stack,” Hykes said. “[This acquisition] gives us a lot more firepower to solve these problems.”

Worked with hosting since 2001, OpenVZ since 2005, and believe Xen, et al. are on their way out due to overhead bloat. I'd expect CoreOS & Docker w/ Kubernetes in tow will eventually become their de-facto replacements.

Docker just purchased Unikernel to go beyond containers, and Google is all in on Kubernetes.

We'll see if I'm right, I could be ranting over Betamax, but I see a trend. We'll see.

nowprovision

@Kris said:
I'd expect CoreOS & Docker w/ Kubernetes in tow will eventually become their de-facto replacements.

Containers are definitely a good bet. Kubernetes is an awesome platform that is extremely well thought out, everything from the logical pod concept, to secrets, to service (and virtual ip proxy), to readiness probes. The current coupling/dependency on either rkt or Docker for images is merely an implementation detail in my opinion, although having access to images and simple forking of dockerhub images is convenient, I wouldn't necessarily marry the success of the two. Similarly with CoreOS, as thin hypervisor OS for kubernetes it probably makes good sense but again once the cluster is up Im not too concerned, and if I use GKE I don't even care (or can care since I dont have raw access to cluster level). tl;dr - definitely kubernetes, docker images are fine/good/plentyful - mutual outlook on larger docker ecosystem, no expectation on coreos

That said, I don't see a problem with KVM/Xen etc.. for those requiring/demanding a VPS or multiple VPSes, utilizing something like Kubernetes just to run a VMs as containers yield little/zero benefits, its not an OpenVZ alternative, if you don't buy into the larger picture of architecturing your site/services/whatever onto a container infrastructure platform then its pretty pointless. Indeed you need nodes (vps/dedicated) to build a kubernetes clusters unless you choose GKE.

impossiblystupid

@Kris said:
Worked with hosting since 2001, OpenVZ since 2005, and believe Xen, et al. are on their way out due to overhead bloat.

I'd say the shifting landscape has more to do with fads than any particular technical strength or weakness. When it gets right down to it, there are any number of equally valid ways to abstract service deployments. Some of us will want the future to be about VPS swarms with resource pools, others will want more from containers, and maybe someone will go off and decide that it's all just about configuration management solutions ([chef|puppet|ansible|vagrant]servers.com all unregistered! :-)

From the low-end perspective, all I really care about is how well the solutions scale down. If I have a very small Rails app, for example, any "grand design" solution for deployment is essentially worthless to me if it costs me Heroku-level prices or more.