New on LowEndTalk? Please Register and read our Community Rules.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
Comments
I think this thread has a lot of drama potential.
I believe in this thread.
Yes, @raindog308, I very much like how @DETio avoids certain questions while competing with Solus, while saying they are better but doing exactly what Solus owners are being accused of
I don't know how many hosts here actually know shit about how development works but missed deadlines? That's a part of our work. Its like predicting the future, you just can't always be right.
But yeah, I can only count @DETio as a reasonable deadline misser. They still have something to show up. Stop demotivating ya'all.
The onapp guy @ditlev on the other hand, either needs to fire his project manager, or needs to hire real developers.
@srvrpro In well run organizations, missed deadlines are a manageable event. They often come from significant lack of planning, feature creep, inadequate resources availability, inadequate competence of programming staffs, and lack of consistent management of burndown lists. It also is based on bundling everything into an all or none release to make the appearance that everything has to be done before its an official release. Most software development is better served by incremental software releases and continuous improvement. So multimonth delays are not necessarily part of software development, but often part of a software development environment without tight fiscal controls.
You said that in October.
I did that In November.
When you have been stable for longer than a year and I can asses the situation, I will consider it, to many things don't add up or inspire any confidence at all.
All of what you've just described only applies when developing well-understood software - that is, software where the scope and complexity are known upfront, there are no big problems to solve, the solutions are well-understood and/or well-documented, and so on. It's effectively typist work for the most part.
When you're working in a more 'pioneering' field - and due to the lack of good documentation or centralized knowledge, this currently also applies to VPS panel development - it's not all that strange to run into sudden delays and missed deadlines, sometimes even big delays. This can happen even in an otherwise well-organized project.
What the fuck are you talking about? VPS panel development is a pioneering field? W T F
And you want us to give you money to develop something?
It would probably help if you read the entire sentence including context, rather than picking out three words and attacking me over those.
Yes, VPS panel development has the same problems as pioneering fields, even if it technically isn't one. Primarily because of the lack of good documentation on the underlying technologies, as well as some other factors (eg. unsolved implementation bugs that still require workarounds even for commonly used virtualization features). It also doesn't help that there's not much by way of good existing implementations.
Something average like SolusVM built on PHP? Probably the easiest way, no pioneering required.Something comparable to Digitalocean's (Ruby) or even Vultr's panel on lets say MEAN stack, it becomes hard to have it all covered in scope.pdf
If you can't write a scope document (subject to amendment in a rational fashion over time) you are just programming from the seat of your pants and welcome to the wonderful world of delays and feature creep. Pioneering for the workarounds - yes I agree. But that is a small subset of the larger project where UI isnt pioneering, basic data movements arent pioneering, documentation isnt pioneering. Creating as project plan isnt pioneering. Those work arounds are the bulk of the risk and time but the rest of the project can be managed to minimize risk and the "pioneering features" can be identified ahead of time and prioritized. But the whole project may appear unique, but its not all pioneering.
Actually, when you buy Solus, aren't you really getting two things:
(1) admin and client panel for managing VPSes
(2) a set of install templates, plus an ongoing subscription for future ones
Leaving aside however current Solus may or may not be...
You mean the templates that where last updated by ones I built and provided to SolusVM months ago for free
12 months? It's been over two years.
This is my opinion because no real competitors exist:
First we need to consider only hosters, for all other entities there are lot alternatives.
For hosters:
Medium/Big companies use OnAPP or VMware or Build in-house panel. They need a rock company that provide 24x7 REAL support. We use OnAPP and it's great, costly of course, but when we need support, no matter that are 3 p.m. or 3 a.m. we got reply within minutes. OnAPP comes with continous new features, now is really stable and features rich. Why searching somethig else? Just to save 1000$ or 2000$/m? No, these savings do not justify it
The rest of the market are little hosters that want pay peanuts and leave the ship to save $1/m. Companies can't have enough revenue from that market.
Simple.
The bottom line becomes, free/cheap/open source, stable, feature rich. pick 2.
Everyone wants their cake and to eat it too.
@matteob thank you for grouping every host here that does not use OnApp in to some ground sucking scum that wants to pay peanuts really insightful of you looking down from your ivory tower in the sky.
I have said more times than enough I am happy to pay more, I begged solusvm to increase the license fee just to continue development and support.
OnApp has HA and support, aside from that it is pretty crap for end users and despite owning it for 2 and a half years they provide no migration path from solusvm which I would have taken.
The thing is everyone says solusvm is cheap, it is not 'that' cheap, it has a cheap entry yes, but after you have been in business a while your paying 600 - 1000 p/month easy, for a product with basic storage, no HA, incredibly slow development and no real support, compare that to whmcs or cPanel, whmcs is much cheaper, cpanel works out much cheaper as well because of customer density.
As control panels go, solusvm is not that cheap.
Please not feel target of my words, i'm talking for the market in general, not case by case.
I'm talking analyzing the numbers, i not have any single host in mind, but you can not deny that most used phrase here is "OnAPP is too costly".
Lot of people here are searching vm for 10$/y then cry if they not get reply within minutes. Same is for the hosters, they want cheap/free panel, but crying if they not have 24x7 support, constant updated etc...
How much is OnApp currently charging to start out?
They seem to have changed their licensing unit model again.
True, but OnApp is like $1000/month minimum, and you pay $10/month per core. Are you OK paying $80/month in licensing for a single node because you have some dual quad core sitting around?
I mean, Proxmox has HA which isn't bad. No shared storage product but you can use ceph, some purpose built iscsi, or whatever other storage platform you want. Modules Factory is like $60 ~ $70 one time for it. There isn't a lot of development going on but that's likely because he's hooked into everything Proxmox provides and he's at their bidding for things. It's ugly, don't get me wrong, but a UI person can likely be had for a few hundred (or you just look through some bootstrap themes and do it up yourself).
Francisco
Consider that i have enterprise plan with > $5000/m... yes, and customers are very happy about all features that they have. You not need to consider only H.A., but also:
Proxmox and OnAPP can't be really compared.
... and much more.
If i remember correctly last newsletter, they raised entry fee to 2000$/m
I've been monitoring OnApp for a long time now, even from the earlier versions (massive audiences on WHT and hosting open days). SANity (integrated storage) was rough but now everything appears to be mature.
Initially the bare minimum package from OnApp was 48 units at $500 USD per month. Price increased to $600 a few months ago for new customers. Now it seems like the 48 unit package is discontinued and minimum is 200 units. (expect onward $2K per month)
Lets consider the pricing:
Budget for 1/4 rack (10U) in Northern Europe (not England):
Approx $400/USD per month.
Bandwidth: Assume 100mbit/s dedicated for $100 USD (unlikely with most quality carriers and at least $2/mbit on that type of commit). Mostly in North EU, 0.6EUR per Mbit on 500mbit/s+
OnApp licensing costs $500 per month (48 units).
Total: 400+500+100=$1000 USD per month.
48 units = physical CPU cores. Use 8 physical core servers. You can have 6 servers to maximise use.
To break-even at $7 margin, then you need to sell 142 machines. If you want any significant return on hardware, then you need sell 2x (284) or 3x (426 machines). By then you will be saturating disk IO and perhaps CPU usage. You may even saturate your bandwidth link too! On the good side, OnApp can definitely handle that many VMs!
You can understand why OnApp is hardly sustainable on Lowendtalk price range. No overselling of disk or RAM in calculations.
Don't forget investment in Dell/HP etc enterprise SAN if needed (5K+).
Personally I think companies like Dediserve are able to serve because they have a share of corporate or business customers - depending on their HA platform for certain services. What such companies may be selling is excess capacity and hence low pricing can be met.
What I am looking for: A cloud orchestra similar to OnApp with pricing around $100-250 USD per month (48 units type or 6 hypervisors). Unfortunately, no body offers that at this time!
Other options:
OpenNebula - hardly known, needs a lot of time to play with storage and other stack configurations. GUI is good and seems to be some plugins available for WHMCS.
Openstack - CONSULTATION and man hours if you have!
I guess you can see why everyone uses SolusVM or Proxmox thesedays. There are no more options
Even less heard of is VMManager, but they seem to work only in Eastern Europe. Posts on LET shows not favorable results (marketing).
After 3 years, I am still waiting for someone to develop a cloud orchestrator with pricing around $100/mo but not more than $300/mo. There are none.
VMWare is not affordable with VSPP licensing.
I wouldn't use Xen and only interested in KVM.
Hyper-V and Windows Azure Pack is expensive for licensing (Win Server)
But lets get serious, from an end user perspective, OnApp offers less features than solusvm, so as an end user why bother using an OnApp host over DO/Vultr, frankly cloud.net is a joke, I have used it, the feature set is shit.
Admin bells and whistles aside, OnApp really does not offer much for the money, happy to get a real comparison chart up if anyone wants to for fun?
Cloud.Net is a separate project. Code is avaiable on github. Most of feature are not present.
You can't compare Cloud.Net with features offered by OnAPP enviroment (on end user side)
We have few remaining nodes on solusvm and use OnAPP, and there is no comparison.
Well lets do a comparison anyway, see if the price tag is really justified.
Pricing for some exotic locations is nice, but overall the feature set is indeed shit. Requests/changes take a "solusvm 2" time to be implemented.
Which shows how actively OnApp develop their products, honestly anyone that thinks OnApp gives a crap about anything but vmware orchestration right now is absolutely kidding themselves.
Maybe they just did the maths and decided buying anything that could possibly threaten them within 3 - 5 years is cheaper than paying a larger team of Dev's to actually stay in front.
So, to be more constructive, if someone is thinking about implementing new control panel (non open source but cheap) which features would you like to see? Basic CRUD for VMs, KVM support, multi host, IP management, what else?
If you'd be writing it that could be very interesting