Metal as a Service
maas.ubuntu.com
maas.ubuntu.com
* Auto-enrollment of nodes was tough to get working.
* Overlay and config management was weak and better handled by Puppet. Though Juju beans were touted as able to handle this.
* Juju 1.13 and maas 1.3 do not support isolated juju environments in the same maas cluster. No idea what version it is at now.
* Juju charms "local provider" is potentially very compelling but wasn't in a workable state.
In the end we abandoned it.
A stack was run on it, however it lacked the polish, support and flexibility required to make it a platform worth investing in.
That is not to say that we thought the idea was bad but that it was nowhere near mature enough for production last year.
Given how Tahr has rolled out, we suspect most development effort was put into the mobile platform and a close integration of MaaS with the cloud in a box.
You're incorrect about that, but the MAAS team did suffer some short staffing last year that left us behind the curve. We've now increased the number of people working on MAAS, which is what allowed us to get large chunks of work – like integrating CloudBase's work for allowing MAAS to provision Windows machines – relatively quickly.
To be clear we are a great fan of Canonical's work and it's too easy complain about the few issues that praise them for the great work that is done every day on the Ubuntu platform.
The bleeding edge of Open Source technology is oft a perilous place to do business.
Truth is that the problems that MAAS is trying to solve are hard, and it's taken us some time to solve them. We're spending a lot of time in this development cycle working on the robustness of the MAAS node lifecycle.
MAAS 1.5 is significantly better than 1.3, and getting stronger all the time. Similarly, Juju 1.18 and 1.19 are leaps ahead of 1.13 – the local provider is now in a workable state and gets used a lot internally.
We at Canonical eat our own dogfood as far as MAAS and Juju are concerned. We're using MAAS extensively in our DCs (though I don't think we use it for 100% of machines; I'm not privy to that knowledge) and the vast majority of our core services are deployed using Juju.
There were a lot of cool working parts such as the Juju-gui and the way that nodes could be tagged for different charms. Though the way that charms deployed at the command line sometimes didn't appear in the gui vexed us.
I like the idea of MaaS on the cloud in a box. A plug and play cloud has its uses though TechOps time is at a premium and rebuilding the stack unlikely.
We ended up investing more time instead into Jenkins and gave up on the distributed computing issue by using 64 core, 512GB RAM servers.
...and focus on Linux on the Desktop... instead of trying to be an "Everything" company. Get the core down solid and making money, then expand into other markets.
Most homes have either a desktop, or a laptop, or both, sometimes multiples of each.
Every business has at least 1 computer (desktop or laptop, sometimes multiples of both).
So there very much so is a lot of money still in desktop.
Right now if you want to purchase a factory Linux computer, your choices are severely limited. It's either System76, or the Dell Sputnik (which for some reason costs more than it's Windows counterpart).
XP just expired. My company has a warehouse of 40 workstations running XP. We're looking into getting new hardware. Guess what... it'll be Windows because there are not many other options to price-shop around and/or get differing levels of support.
There is indeed money in Linux Desktop.
On top of that, personal computer sales are stagnating and tablet and mobile markets are growing; if Ubuntu really wanted to get on the money train, that's probably a better place to go. There's no money in Linux Desktop any more than there's money in Windows Desktop, which is to say not much at all.
The only PC-related industry that might be worth focussing on is laptops or tablets-as-laptops (e.g. Surface), but there's a huge amount of engineering work to be done on hardware support, battery life, etc., and most of the Windows 8 tablet hardware is garbage. Without creating their own Surface-type product (and having an actual market for it), it would be difficult for them to control enough to ensure proper support.
Compare that to servers, where companies will shell out $3-10k per server for new servers (or hundreds of dollars a month). If you can add in some kind of premium support, management system, cloud something, etc. and charge a minimal amount, the money would be significant. Heck, give Amazon customers the option to opt in to a $1/mo support upgrade and that would be a decent income stream from people running their own servers who aren't highly technical.
Out of all the Linux vendor's and distro's, Canonical is the closest at achieving common-place... yet, they have stopped short at such a critical time.
Of course tablets are popular, so are phones. Both markets are more fierce than desktops due to massively low price-points and some companies (looking at you google) willing to sell their tablets at a loss.
Companies spend $3-10K on server hardware, not the OS. When you buy big-iron servers, it's actually very common-place to get a Linux OS already installed, or no OS so you can install your preferred distro. Server Software (the OS) support contracts absolutely make money... but so does desktop software support contracts.
Your assertion that desktop industry is not worth focusing on is not true. Tablets and phones are getting more popular, but desktops and laptops are not going anywhere anytime soon.
They want to be in every consumer market, but they're too late to them all. Desktops are shrinking. They're having trouble finding a phone manufacturer that will do hardware for them. Probably the same with tablets. I can't even begin to imagine what all is involved in television. That's the market they want into that still hasn't been totally cracked by Microsoft, Apple, Google, or Amazon. And that fact alone should probably tell you that the odds of Canonical cracking it are not fantastic.
I'd love for Ubuntu to become huge on desktops and make them tons of money. And maybe it will happen and I'm just not seeing the big picture. But where you see every business running at least 1 computer and consider that an opportunity, I unfortunately see it as a missed opportunity. That's a computer they already bought, and if they're going to get rid of it and switch to something else what makes you think they would choose Ubuntu (something they've never heard of) rather than Windows 8.1 or an iPad?
Edit: But your point totally stands. They need more sources of revenue besides desktop linux.
But it doesn't really matter. Canonical is drifting from what could be great. The average user does not need tech support to use a modern linux desktop (have you tried recently? they have come a long long way even from 3 years ago). At a time when XP is EOL'ed, and people are looking for anything not Windows 8 -- Linux should be answering that call, but instead Canonical is waffling again. There is absolutely a void in the market.
Let's not make this an argument. Time will tell, as does history. Canonical has fallen short on execution on almost every major project thus far. Maybe this one won't repeat history.
Published stats [1] show that Ubuntu dominates cloud deployments.
Also, he's talking about deployments on Digital Ocean, ie. VM's. I'm talking about what actually makes the "cloud", ie. the "cloud". The "cloud" is built out of RHEL/Centos for the most part. Almost all the major "cloud" infrastructures have been built with RHEL/Centos. We can speculate why... but Red Hat has built a $1 billion+ revenue per year company out of "cloud" and enterprise support contracts. (because they were there first, and they focused on it... and after it became profitable, then they finally branches out into other markets.)
The problem with Canonical IMO is they try to get into too many markets which ends up resulting in split finite resources. "Jack of all trades, master of none".
For one single data point, that's exactly why I use Ubuntu server for all my cloud deployments - because I use Ubuntu desktop and know the Ubuntu/Debian ecosystem really well.
Nothing wrong with that current business model, any problems are more a matter of execution and competing against - as you point out - heavily entrenched first movers. But focusing on desktop won't save them, especially since desktop is saturated and already peaked. Cloud and mobile is where the growth is now.
I'd be surprised if they were the only ones.
DigitalOcean, Amazon, Microsoft(!) all list Ubuntu as the most popular VM for linux servers.
Can you point to stats or facts about RHEL deployments backing your point of view? I'm genuinely curious.
First, to clear some things up, I'll avoid conflating traditional enterprise servers and hosted servers with the on-demand, pay-as-you-use cloud model. The later is what I mean by "Cloud", not just hosted servers.
To my knowledge Red Hat only holds a lead in server OS installations with paid support, and that's almost all in the former model (non-cloud, for me). Incidentally that's also the stat you'll see when companies publish reports about Linux's market share in general - they often just look at who's buying Linux. It's hard to get a real picture of what people actually use when you include unpaid Linux usage (which is most Linux usage), but the best data I know of is created by scanning web hosts, and suggest that Linux server run in order Debian, Ubuntu, CentOS and RedHat:
http://w3techs.com/technologies/details/os-linux/all/all
Red Hat's sudden obsession with OpenStack in the last 12 months, the move to take over CentOS and the purchase of InkTank are reflections that Red Hat are struggling to get in that game and are prepared to spend big to do so.
Why are they struggling so hard? Because they're frightened. The steady, "we want everything to stay the same for 10 years" model they've been serving so well is seriously threatened and they need to be in the new market to continue their success. Businesses love the idea of the cloud (if not yet the reality). Read what companies say about the cloud - they'll talk about it in terms of flexibility, and opportunity, but what they really love is that it means that you can offload the genuinely generic work of IT infrastructure management. This reduces the reliance on staff and consultancy costs, which are the most expensive aspect of corporate IT (hint, want to stay employed, get with the program, fast!). It also means that whether you're using Red Hat or Ubuntu, you get to negotiate the fee for OS level support via a behemoth like HP, Amazon, Microsoft or Google, or if you're a huge company with a private cloud, you centralise all that and do it in bulk. The same goes for the physical space for hosting.
Now Amazon were an existing Red Hat customer and so their servers were already running that when they started on this public cloud adventure - which afterall was a clever way to reuse their existing investment in infrastructure and personnel outside peak load time. That's not a big win for Red Hat, but there's growth of their existing business for sure (who knows what kind of deal Amazon got from them?). Still, the utilisation of that cloud is the bigger chunk of new business there, and it's Ubuntu winning the majority of it.
Now spread out to the nascent cloud market as a whole, sure Amazon are a huge public cloud with the bulk of utilisation, but there are many others, new ones all the time. Additionally and many, many private clouds that are being created those are the opportunities for new revenue in the hosting market. Almost all of them are going with OpenStack and 55% of OpenStack deployments are Ubuntu right now (I've seen Red Hat running at less than 1/10th of that business. I'd expect Red Hat to increase that number a lot in the next couple of years (they'll convert some existing customers for sure), but don't be fooled by the recent, aggressive press releases from Red Hat - they're fighting to get a foothold in this market, not controlling it by any means.
Straying into my opinion. I want Canonical to do very well. As I mentioned, they are the closest anyone has ever been to becoming a household name with their Linux desktop (they are sooo freaking close!). But the problem I see with Canonical is the company seems to have a very bad case of ADHD. Canonical tries to do everything. Phone, tablet, desktop, server, windowing manager, init system, cloud, SaaS, enterprise support, package manager, weird-NUC-box-things, app-store, etc etc etc. This is far too much for a company that is barely staying afloat (they posted an almost $22 million loss for the same fy; about 1/3 of their total revenue lost). In most of the markets they dabble in, they show up late (sometimes very late), can't get a strong foothold, and will or have been shoved out. Shuttleworth told theguardian in August of 2013 that Canonical wanted to capture 25% of the mobile market in the US and Asia... an absurdly unrealistic goal (to this day Apple has only managed 21% market share according to idc.com) Mobile may be "hot" right now, but they are a no-name brand in that market and they are attempting to go up against companies that can throw 10's of billions of dollars at products, and even take 10's of billions of dollars of losses just to ensure they remain dominant. The ship on mobile has sailed.
I applaud their engineering efforts, and I'm sure Canonical employees do very good, and very hard work. But they need to reel it back in and focus. The company can't do everything Mark thinks might be cool when he wakes up in the morning. It needs focus. If Canonical were a startup, and Mark didn't have deep pockets, it would have failed many times over (Mark has acknowledged this several times citing his commitment to Canonical and willingness to re-invest personal capital as necessary).
Back to Desktop... there is money to be made there. Several OEM's are struggling, consumers don't like the current options (Windows 8 or OSX). There is a void to be filled, even if it's only 10% of the market, that could result in a huge revenue increase for Canonical. Yes, it's difficult to get into the hardware game, but it could be done. To make revenue, sell support licenses/contracts to OEM's... use the Microsoft model for desktop support. Why is System76 supporting Ubuntu themselves? Canonical isn't getting anything out of that... and supporting the OS costs System76 money no doubt, why not help take it off their hands? The Dell Sputnik? Dell is supporting it directly, and charging consumers more for the privilege of having a "free os" over a Windows license. Over the next 2-5 years Ubuntu could become a common household name, and be no less-normal than installing Windows on your pc. There supporting software is there, the desktop is mature. It just needs someone to guide it in.
If Canonical does not want to do Desktop, well I think that's a mistake, but fine. So they should pick one market and do it very well. Don't split-up your resources; there is only a finite amount.
So Red Hat could buy Canonical? I actually think the resulting company would have an awesome market position (but a terrible challenge to unify their tech-worlds). I think the linux world would be less rich and exciting as a result though. Still it won't happen, because it rather depends on the shareholders of Canonical don't you think? I suspect it wouldn't be as easy as all that to convince them (him!). I saw an article recently suggesting Microsoft should buy Canonical. An even worse idea IMHO - Ubuntu users wouldn't go for that. They might even see Red Hat as a bit too corporate.
Re: Canonical's financials. Which do you want, bigger revenues or profitability? Canonical could be profitable with it's current revenue levels, if it stopped investing in new things (an idea you seem to support, but be careful killing the cloud products if you're after a profit). They could also have tried to make a lot of money from limited utilisation with mandatory support, at a higher price, a la Red Hat - but how would they have competed with Red Hat if they used exactly the same model? Someone, somewhere must have a reason not to do that, don't you think?
Here's another question. Why hasn't Mark Shuttleworth just given up and walked away after 10 years of not making a profit? There's a tendency to look from the outside and assume that this must be a terrible business decision, but I find it dangerous to make assumptions about the motivations and strategies of others, especially those with a track record of success. You might be mistaking ambition for delusion. If you're wondering about Mr Shuttleworth's understanding of the current market place check out these two headlines:
1. http://www.inktank.com/ceph/mark-shuttleworth-believes-ceph-...
2. http://techcrunch.com/2014/04/30/red-hat-buys-ceph-provider-...
Guess what - that means Red Hat just gave Mark Shuttleworth a bunch more cash for his share of InkTank, at a much higher market valuation than he bought them at. Foolish business man that he is. Maybe it's dumb luck. Or maybe Mark Shuttleworth saw the potential and invested in it at a relatively early stage, and even got the thing into Ubuntu and supported it so that customers could get to grips with it painlessly. Two years later InkTank had a decent customer list and Red Hat acquired it to try and gain a foothold (and some control) in the OpenStack market.
Now, the desktop. No one has ever said Canonical isn't going to do Desktop, just that you shouldn't necessarily focus on it as the main source of potential revenue for Canonical. What about those finite resources? Canonical has more resources today than it did 5 years ago, you have to do something with those millions of revenue to stay in the red! ;-) It may still be a smallish company, but it's not the case that Canonical sacrifices one thing to do the other. You'll notices that there are new partners shipping Ubuntu machines , especially in the developing nations - there's a reason for that and it wouldn't be happening if Canonical didn't care about it's desktop. Is Canonical overstretched? Well, it could be, but that's the fight, that's the nature of every start-up, every business perhaps. There are probably people at Apple right now moaning that they don't have enough resources.
Linux on the desktop is a tough sell to non-technical users (who don't need support, etc.), but technical users operating server clusters is big money.
They've demonstrated they do not have what it takes to make desktop net profitable for them.
They make most of their money by selling support for their cloud stuff — this is what they should focus on. Desktop has done nothing but lose them money.
Juju supports many backends, one of which is MaaS.
> * Auto-enrollment of nodes was tough to get working. > * Overlay and config management was weak and better handled by Puppet. Though Juju beans were touted as able to handle this.
If you have any detail to share on these I'd love to hear more. The first point has long been a sore point in MAAS and we're doing a lot of work this cycle to improve auto-registration and dealing with BMC hardware.
> * Juju 1.13 and maas 1.3 do not support isolated juju environments in the same maas cluster. No idea what version it is at now.
You can definitely do that now -- all you need to do is to set up separate users in MAAS and use different credentials for each of those. We've used this recently in the field for demonstrations (to take advantage of limited hardware)
> * Juju charms "local provider" is potentially very compelling but wasn't in a workable state.
This also works today (and can be used independently of MAAS) and as of 1.18 is in great shape; I'm personally supporting a few companies using it in production.
Virtme: a very promising start to get a proof of concept up and running -- sadly it appears to be abandoned?:
https://manage.jujucharms.com/~virtual-maasers/precise/virtm... http://javacruft.wordpress.com/2013/06/25/virtme/
A (partial) list of manual steps that try to achieve the same (with the goal of ending up with an openstack poc):
http://www.teale.de/tealeg/computing/cloud/kvm_maas_juju_ope...
YMMV etc.
Cool cluster btw, it'd be neat to see you blog about using MAAS on it.
There's something quite magical about using Juju and MAAS on the Orange Boxes and seeing the LEDs light up as you scale out a service (e.g. juju add-unit -n 5 nova-compute).
It eliminates the virtualization host layer and, most importantly, interaction from other guest VM machines.
OpenStack 'can', but in practice the baremetal nova driver[0] has many issues, and the newer OpenStack Ironic Project[1] is not yet ready for production workloads.
Additionally, both baremetal and OpenStack Ironic have a focus on 'triple-O', or OpenStack on OpenStack[2] use cases -- that is using these projects to bootstrap an 'undercloud', for the hardware to be used by a higher level virtualized cloud.
[0] - https://wiki.openstack.org/wiki/Baremetal
I was totally convinced that Amazon had baremetal instances available, but now I cannot find anything, so that could probably be a figment of my imagination.
Well, I'm sure they have issues but what gave you the impression they are focused on TripleO use cases?
As a TripleO dev I just use them.
Compared to more 'standalone' softwares like MAAS, these tools are integrated already into OpenStack, allowing you to deploy images on either virtual instances or baremetal without adding more software layers in charge of the images hosting, authentication, dhcp management and what not.
Might as well call it physicalization. :)
The installation was straight forward and easy.
I currently manage one VPS that hosts multiple sites/apps that I work on, and it's not the worst thing ever, but it could be easier. Right now I still SSH in and restart services by hand, but a web interface would make that much easier to manage.
If they allow for extensible plugins, we could see an explosion of nagios-like functionality being developed, as well as even better free server monitoring resources.
Really looking forward to seeing this progress
A similar situation happened when Apple adopted the name Launchpad for one of their services.
* why do they require a "cluster controller" for each subnet (or vlan)? This seems like a waste of resources and added complexity. Is this a limitation because it's easier to set up for sysadmins than to request for a DHCP relay?
* how bad is the security? It seems that you approve clusters interfaces but MAC spoofing could do a lot of harm here.
Edit: I saw that the doc tells you to add network interfaces to manage multiple clusters from a single controller. Assuming you do 802.1q this works but what about multi-dc setups? Isn't this more complicated than a simple relay?
https://ubuhulk.wordpress.com/2014/03/10/canonicals-office-o...
FTFY :)
But you do raise an interesting point, which is that all the MDCs have their own homegrown infrastructure (which incidentally they will tell you they love to hate, or vice versa). That is definitely true, and they are not MAAS target customers. But it's also true that they are not the only places where you will encounter large deployments of servers, and if you believe that at least a segment of the server market will go towards more/many nodes (a.k.a hyperscale), then having something to reliably deploy and manage a bare-metal cluster is definitely essential.