Why Cloudmin

Hello. First of all I’ve always appreciated your honesty and good will so I trust this won’t be different. But I have to ask, cause besides many things I’m a noob in cloud computing and Linux virtualization:

-why should we choose Cloudmin (except the web interface and the obvious differences) over another free/open source software like Convirt, Proxmox, Enomalism etc?

It’s not about the new KVM trend brought by Red Hat & CentOS, kind of a minus, but you will surely fully support the technology in the future, i know, it’s about the annual renewal prices somehow scary for me and the missing terms of comparison with the mentioned.

I will try all of them this week, but exactly yours, the one I probably want and for wich I have to pay for - I can’t.

For example, let me tell you what convinced me that Virtualmin Pro is for us:

  1. Webmin and your history (this counts already for Cloudmin);
  2. The demo (this one is pretty stupid for Cloudmin I guess; but what about a trial? wouldn’t be more easy for you to have fast expiring licences than to set up a demo?);
  3. Virtualmin GPL - running for years on one of our machines; one of my partners cried when we stopped an old Athlon 3700+/3GB with Vmin GPL :slight_smile: - so… an addiction;
  4. The comparison table between the hosting panels and my experience with the others - a how to doesn’t explain plainly what’s your edge/advantage, isn’t it?

And do you plan an Open Source implementation? Trust me, in this question, it’s not about the money, but about the new technologies, growing habits, and defining a way of doing things. And a decision that will affect my evolution as an IT enthusiast and my small business.

This is also the kind of info we need https://www.virtualmin.com/node/10363 You really should answer those last questions.

And you have to work on your marketing :)) and your success, because it is one! I see roughly the same questions everywhere in the forum :))

Thank you.

Thanks for the thoughtful post.

So, I’ll talk a bit about our goals with Cloudmin, which will probably make it clear why we’re pretty dramatically different from most of the products in this space (though Enomalism seems to have similar goals, in many regards, and so can be viewed as a competitor to what we’re doing).

Our primary focus is on building shareable cloud computing resources, to be used by hosting providers (and enterprise with large data centers and diverse groups of users). This goal involves a lot of tasks that are not addressed by most of the current crop of products in any meaningful way…things like delegation to somewhat untrusted (though generally not completely anonymous) users. As in the case of shared hosting, there are a bunch of additional requirements that hosting providers wanting to offer cloud computing services have that normal users of virtualization simply never think about or care about.

So, from that description of our focus, it should be apparent what Cloudmin is not for: Single users with a small number of virtual machines to manage. If you aren’t sharing your cloud computing resources, you don’t have more than one host machine, and you don’t need a way to keep up with usage in a fine-grained manner, Cloudmin is probably overkill. If, on the other hand, you are a hosting provider that wants to begin offering cloud computing services, or an enterprise with an existing data center and want to begin building a private cloud, Cloudmin might be a good part of the solution, and we believe our past experience with Virtualmin and Cloudmin and the hosting industry makes us uniquely qualified as effective partners.

Now, on to your specific questions…

KVM support. It’s coming. But, you should be aware that Xen is still better suited for the kinds of deployments Cloudmin is best suited for (and some people think OpenVZ is even better because of its lower resource usage, but we tend to believe the tradeoffs are too many and too large to be worth saving a little RAM). We haven’t actually begun writing the code yet, but I suspect it will be supported within the next month or so (new virtualization types are pretty easy to add).

Trial. Also coming. Very soon, actually. I’m hoping to finish up the website support for free trials this week. Jamie will need to add some code to Cloudmin and the license manager, but I suspect it will come in the next release of Cloudmin (2-3 weeks away).

Open Source version. Very likely, but I haven’t yet been able to figure out how to make it pay for itself the way Virtualmin Professional pays for Virtualmin GPL. What do you think we can charge (enough) for while still offering a usefully functional Open Source version? We’re pretty happy with the balance in Virtualmin Professional vs. GPL, but there doesn’t seem to be as obvious a breakdown for Cloudmin…since it is almost entirely useful for commercial entities. The Cloudmin banner will soon cover a number of products, designed to help folks build and manage cloud services for sale or delegation to others; we need to be able to pay for their development while still being true to our Open Source roots (as I’m sure you can imagine, we like Open Source; we’ve both been building Open Source software for over 12 years, pretty much our entire professional careers, and we don’t intend to stop now). So, we’re trying to figure out the right balance. Unfortunately, “support” is not a profitable business model for most kinds of software, and it also puts the incentives in the wrong places for the kind of software and business we want to build. Plugins and addons might be a way forward for an Open Source Cloudmin, or perhaps a Professional version targeted to making money, as with Virtualmin.

I’ll take a brief little detour to talk about the current state of the market for virtualization management. It’s very early in the story…all of the products that exist (including our own) are extremely simple products; most have less than a year of real development on them (some developers are faster than others, and Jamie has closed the gap with the few folks who got started before we did, so we’re close to parity with everything that does exist our target market). It’s extremely easy to build a product with the minimum functionality that you see in most of the products in this space. There is no rocket surgery happening in this space yet; the hard parts that are being solved are being solved at a lower layer than the management tools (the Xen kernel developers have solved some really hard problems; likewise KVM and the Zones team on Solaris, and to a lesser degree the OpenVZ folks). We’re just now starting on some of the really hard problems (scalable reliable shareable storage, database management, queue and response), and we have some of the problems already partly solved by existing Webmin code (status monitoring, notifications, RPC, API, etc.). I believe pretty strongly we will get many of those problems solved before our competitors, because we have such an advantage being based on Webmin, and having Jamie leading the charge (I’ve said it many times in the past, but I’ll repeat it here: Jamie is as good as ten good developers, and when he sits down to code, I’ve never seen anyone work faster).

And, yes, we are working on our marketing, and making the right information easier to find. The Cloudmin docs have been completed and are now online at http://www.virtualmin.com/documentation/cloudmin and we’ll be posting information and tutorial videos as soon as I wrap up the production on those. The Cloudmin landing page is also underway and will be published soon. We’re in the position of having an existing product that is still growing in popularity and profitability (Virtualmin) that makes it hard to just drop everything and focus on Cloudmin…from the perspective of people that only care about Cloudmin, our website seems close to useless, because it is focused on informing people about Virtualmin. To people that don’t care about (or already know about) Virtualmin, it’s just noise and in the way. So, I’m working on getting a clear path for information about Cloudmin starting from the very first page of our site. Should be getting better over the next couple of weeks.

I’ll also make note of some seemingly unrelated stuff: We’ve asked Eric to spend some more time working with us and to expand the projects he works on; and we’re hiring at least one additional person in the coming weeks. This will allow me to focus on getting the website and our marketing in order, while our support and development can continue without distraction. We also have a couple of third party apps planned that will make our products cooler and more useful for a large number of folks (I won’t name them specifically, but iPhone/Android users, as well as Drupal users, will find these additional projects useful).

Maaan, now that’s an answer! I’ll get Cloudmin when it will handle KVM, I was already focused on your product; I tend to panic when Red Hat & CentOS kicks off something, sort to speak. Hope you will still have your introductory offer :slight_smile:

Speaking for myself I’ll opt for KVM. It seems that Xen always hits some problem, everything is unstable; it’s me I know, but I’m hopeless.

Enomalism is a very cool tool indeed, and you’re right this is your main competitor, forget the others. But I think I’ll put my option where Virtualmin is… it’s somehow logical I guess.

But you’re wrong somewhere: I don’t think that the cloud it’s not for the average guy/biz, though I understand that someone has to have plenty of resources in order to sell them - the problems are pretty much the same if you manage alone 3 servers with a few virtual machines each, my case, or 13 with a dozen of them each. And lots of domains. It’s kind of a threshold issue - when you’re reaching your limit it doesn’t matter how far the problem goes in itself, productivity is gone, and you have to solve it for 1+threshold or 1000+threshold (probably bad English?).

After the last post I tested the hell out of cloud solutions (I’m sick of installing OS-es :slight_smile: and I know now that the Cloudmin has the right tools for the job (from the docs only), especially the web interface, live migration, managing pools of physical machines, all those great things with Virtualmin, and easy, regular backups which don’t ask me why is kind of a problem - because it’s one of the greatest things in virtualization and one of my issues: it’s easier to get a new machine running if it fails by putting an OS and restoring a VM image already set up, than setting up a hole new standard server and restoring the domains.

Glad to hear that you’re on your way to make things simpler for us, if it’s a technical problem or a marketing/information one. You always did that.

Thanks - I’ll keep an eye on Cloudmin.

PS: Convirt it’s more like a replacement tool for the virtual manger. It has a lot of features, snapshots being probably the most important. But I don’t like the way that Convirt console works - that mouse is really annoying - and I don’t think it’s ready for production use.