AWS hell no: Can Microsoft Azure sales beat Amazon's cloud?

There's a prospect coming into view of Microsoft’s Azure cloud offering passing Amazon's AWS in cloud revenues. Fancy that. Azure grew its commercial cloud services revenue 147 per cent on an annual basis in the second quarter of 2014, and has an annualised run rate of $4.4bn. So says financial analyst Aaron Rakers, the MD of …

[Richard Taylor 2]

Loong term view

It might not be popular but do you put your resources into a company that is profitable - or one that is not (a lovely subsidy) but is building market share and may come back to claim 'back rent' once you are dependent on them?

Not that I think either are a good long term business prospect for many users..... (concerned with choice, security, performance ect ect ect)

[Anonymous Coward]

Re: Loong term view

Quite - AWS have yet to make a pound - Amazon could drop them tomorrow or sell them off. Azure is much easier to use and consequentially has a much lower cost of entry.

I think that AWS will likely end up with mostly 'Digital' or 'Web' shops where the large overheads of having to run a 'Dev / Ops' type operation and using highly labour intensive to maintain deployment / automation tools like Chef and Puppet are a cost of doing business in that space. Other enterprises are tending to head to Azure or similar platform strategies in the private / hybrid cloud space.

[Anonymous Coward]

Re: Loong term view

"Azure has a much lower cost of entry and is much easier to use."

Not sure about the cost part, but by which metric/from which environment you find Azure easier to use? If you're using Azure from a Microsoft ecosystem, then likely it is easier, but what about the rest of environments? (such as Linux, Mac)

Always have the impression that Azure tries to grab the market share of small and medium sized Windows users where the integration is valuable, but for the software development business that are not 100% MS centered, is that the case?

[WP7Mango]

Re: Loong term view

Azure IS easier to use, period.

Azure is also fully supports Linux, Oracle, and other non-Microsoft technologies including iOS and Android for simplified back-end service integration.

[Arion]

Re: Loong term view

You know, capitalising 'is', and putting an American cliche at the end of your sentence, doesn't make it any more persuasive.

[Anonymous Coward]

Re: Loong term view

Oh, I wondered what period he was talking about. You just made me realise that's a colonies anacronism for a fullstop.

[WP7Mango]

Re: Loong term view

"You know, capitalising 'is', and putting an American cliche at the end of your sentence, doesn't make it any more persuasive."

You're confusing me with someone who gives a shit.

[Trevor_Pott]

Re: Loong term view

No, we accidentally confused you with someone who had wits.

[Semaj]

Re: Loong term view

"Azure is easier to use"

I use Azure every day and can confirm that it most certainly is NOT easier to use than AWS. At least from a development and deployment point of view.

I'll list a few off the top of my head examples of where Azure makes things more awkward than AWS (or an on-premises setup, which is basically the same as AWS) I'm talking from a .net perspective, which in theory should make things easier so I dread to think what it's like for non .net folks.

- Wrappers. To deploy to Azure you need to wrap your projects in a special .cloud project. These frequently become corrupt (especially when upgrading) and have lots of config XML duplication. Very awkward to merge. (You don't have to use these if all you are making is 1 website but for anything remotely interesting you have to use this mechanism)

- Deployment. Is SLOW. You can deploy from Visual Studio, which packages up your application then sends it up to Azure, which will then makes a whole new VM, installs what it needs then allows you access. This can take anything from between 5 and 45 minutes depending on the size of your application and which way the wind's blowing. Note that getting the deployment set up in the first place is a pain because you need to mess around with storage locations and other auth stuff. Oh and if you just want to make a minor CSS tweak (for example) to a wrapped website you need to do a whole new redeployment because any changes you make via RDP aren't guaranteed to remain.

- Config. With Azure, they've thrown the .settings stuff in the bin and implemented a fudge so you can store your settings in the .cloud project and access them through a proprietary MS class. This does fall back to web.config or app.config if the setting isn't found but it throws and catches an exception to do it. Also it duplicates config, meaning it's easy for stuff to get missed out. Oh and it still takes ~5 minutes to make any config change via the Azure front end (it must do a soft reboot or something).

- Waste. For every project in a .cloud wrapper Azure will create a VM, eating into your cores limit.

- Backup. Backups (of Azure SQL) are still practically impossible to get set up in a reliable, automated, on and off-site way.

- Azure SQL. Is just generally awful. Missing features (freetext search, custom CLR functions, scheduling), difficult to diagnose when things are going wrong.

- Scheduling. Is only available via creation of a mobile app and even then it's only able to call a web service so you still need a whole instance per (different) scheduled action.

- Remote access (RDP). Needs to be enabled specifically from the first deployment or you need to redeploy with the various settings enabled. Makes debugging a failing service very difficult. Oh and you can't do an in-place swap of VMs while turning on RDP because it creates another endpoint and you can't swap in that situation. That means downtime, of however long it takes it to deploy.

- Diagnostics. Aren't provided by MS at all. The best software seems to be "Azure Management Studio", which is OK if a bit flaky (I put this down to the services they are calling rather than their software). But to get anything useful you have to mess around with more settings, pepper your code with more Azure only MS classes (Azure Diagnostics), turn around 3 times and stand on your head.

...That's all that immediately spring to mind.

[Gamberoni]

Re: Loong term view

I heard about one you've missed out - although it's not strictly to do with deployment.

I understand that Microsoft will always patch your environment with the latest fixes - whether you want them or not and regardless of whether it breaks what you already have. No rollback either! For me that makes it a deal breaker.

[Anonymous Coward]

Re: Loong term view

"understand that Microsoft will always patch your environment with the latest fixes - whether you want them or not and regardless of whether it breaks what you already have. No rollback either! For me that makes it a deal breaker."

Nope - it's up to you to decide on patching policy for your VMs. Just the same as if you host the VMs in your own datacentre.

[Semaj]

Re: Loong term view

They do but they don't push out major version upgrades. They do deprecate them though.

[Anonymous Coward]

Re: Loong term view

"use Azure every day and can confirm that it most certainly is NOT easier to use than AWS. At least from a development and deployment point of view."

I manage and implement deployment / framework / orchestration systems for both AWS and Azure - Azure is hands own easier to manage implement and maintain as infrastructure and as a platform.

Most of your points seem to be related to using Azuare as PaaS - and very specifically to .Net development - and your lack of understanding of the platform, and that this behaviour is by design - for instance your comment about manual configuration changes not being permanent to systems that are designed not to be stateful!. If you want the behaviour of IaaS then select IaaS solutions and provision and manage the appropriate VM resources. For instance if Azure SQL isnt good enough, then install SQL Server on a VM.

SQL Azure has very powerful, straight forward and simple to use 'offsite' on 'onsite' backup options, including Database Copy + Manual Export, Active Geo-Replication and Point in Time Restore within the last 35 days. Exports can be schduled, and Active Geo replication and the Point in Time restore options can be fully automatic / self service. See for instance http://msdn.microsoft.com/en-us/library/hh335292.aspx#automate and http://msdn.microsoft.com/en-us/library/jj650016.aspx

RDP being disabled by default is a good thing. You are basically complaining about your lack of planing and your inability to use Powershell or other more suitable tools to get the information that you need...

[Semaj]

Re: Loong term view

Well sticking stuff on a VM would certainly work but that kind of defeats the purpose of Azure's offering IMO. We may as well (and should) have gone with AWS if we were going to do that. Besides, they've only been offering actual VMs for a couple of years and our stuff was already on Azure by then.

As for the bacpac option, they don't include the transaction log so aren't fit for purpose for most companies who would need a real backup strategy for a legal requirement to minimize risk.

And of course RDP being disabled by default is a good thing but you shouldn't have to take down your live service to enable it. And you shouldn't need to enable it because Azure should provide access to real diagnostic tools from the admin console.

[Anonymous Coward]

Re: Loong term view

"Well sticking stuff on a VM would certainly work but that kind of defeats the purpose of Azure's offering IMO"

No, Azure offers a CHOICE of Iaas or PaaS. That's by design.

"As for the bacpac option, they don't include the transaction log so aren't fit for purpose "

They dont include the transaction log because the managed service already includes point in time restore capabilities. If you do an export then the database will be checkpointed and transactionally consistent at the point of time of the export.

"And of course RDP being disabled by default is a good thing but you shouldn't have to take down your live service to enable it."

Yes you should if you are using PaaS - PaaS systems are not stateful and are not designed to have the OS config changed on the fly. If you want that flexibility - and the overheads that go with it - then start using IaaS and manage your own VMs.

[Trevor_Pott]

Re: Loong term view

"You are basically complaining about your lack of planing and your inability to use Powershell or other more suitable tools to get the information that you need..."

Yep, you definitely work for Microsoft. Condescension and blaming the user as SOP. Bravo.

[Anonymous Coward]

Re: Loong term view

"Yep, you definitely work for Microsoft. Condescension and blaming the user as SOP. Bravo."

For a well known FTSE 100 actually other than Microsoft, but nice try.

[Trevor_Pott]

Re: Loong term view

First of all: I don't believe that you work anywhere other than Microsoft. Secondly, even if you did work for some other big enterprise, does it make you feel special to be a disposable faceless drone amongst a sea of similar "resources"?

Do you somehow feel superior to others becasue of the size of the company you work for? Does the opportunity to specialise ever more narrowly, knowing more and more about less and less until you know absolutely everything about absolutely nothing make you feel like you somehow know everything?

You're a coward. A faceless, nameless, useless coward. You vomit assertions and tautologies and won't even put your name to your befuddled ravings. There's no reason to accept or believe you are anything other than a paid shill waving the banner of their paycheque in desperation.

Your livelihood very obviously depends on you believing Microsoft uber alles. Mine doesn't. Mine also doesn't depend on seeing any technology vendor succeed, or fail. I get paid the same regardless of who wins or loses because my job is to discuss technology as a whole, not to give a bent fuck who "wins".

We'll let the readers decide whether or not faceless assertions and transparent tautologies hidden behind a veil of anonymity represent sound technological - or business - advice.

[Anonymous Coward]

Re: Loong term view

"First of all: I don't believe that you work anywhere other than Microsoft"

Since when did anyone care what you believe?

"Secondly, even if you did work for some other big enterprise, does it make you feel special to be a disposable faceless drone amongst a sea of similar "resources"?"

Nope - i'm not - im fairly high up the tree. It is you that are far more the commodity resource. :-)

[Trevor_Pott]

Re: Loong term view

"Since when did anyone care what you believe?"

You, evidently, since you can't stop yourself from trying to convince me of your lies.

Oh look, an anonymous coward asserts, and asserts, and asserts some more. Pretends he's a big shot. You obviously are just the bees fucking knees. You must be really important for you to spend so many hours a week on here pimping Microsoft's party line, especially if you don't work for them. You're so important that it just doesn't matter that you don't put that time into your job, or the sanity of having a life, or really anything better than repeatedly bullshitting commenters on a forum on the internet.

Oh yeah, I am thuroughly convinced that you're an indispensable big shot who just couldn't care less what I think. Good job. I bow in awe.

[Anonymous Coward]

Re: Loong term view

"Always have the impression that Azure tries to grab the market share of small and medium sized Windows users where the integration is valuable"

I would say that Azure is primarily focused on large and enterprise sized customers. SMEs often don't have the resouces to develop cloud integration. Not that Azure isn't sometimes a good solution for smaller companies, but I don't think that's the primary focus at the moment.

"but by which metric/from which environment you find Azure easier to use?"

The vast majority of standard cloud / infrastructure / SaaS / IaaS processes that you might want to do are easier and simpler to automate in Azure and the integration is way better to standard corporate systems like Active Directory. From personal experience of both.

"but for the software development business that are not 100% MS centered, is that the case"

I think so with the various Linux support options that Azure now has. And full Azure support from Oracle, Sap, Hadoop HDInsight, etc. etc.

[Anonymous Coward]

Re: Loong term view

And Azure has support for / from Docker Libswarm, Python, Ruby, Node.js, Java, Kubernetes, Packer.io and OpenNebula

[Anonymous Coward]

Re: Loong term view

Thanks everyone who has taken the time to comment. Worth noting that the views of Azure being easier seems to be split like in a Linux vs. Windows debate: some say Azure is much easier, others say it is not, very few in the middle saying the -usually resembling more balanced and reasonable. "well, it depends"

Note to downvoters: the original comment was not reflecting any opinion on Azure, it was merely asking a question. If you feel that asking why something is seen as "better" or "easier" is an attack rather than an opportunity to show off your preferred platform strengths, you clearly need marketing 101 lessons.

[Trevor_Pott]

Re: Loong term view

"you clearly need marketing 101 lessons."

If you're in marketing, why the hell are you on these forums? These are for technologists, not shills. You're a bad person and you should feel bad.

[Anonymous Coward]

Is Office 365 a separate product from "Azure"

Just wondering if those revenue increases come just from Office classic license conversions to 365....

[Anonymous Coward]

Re: Is Office 365 a separate product from "Azure"

I expect it's partly in the same pool. Not unreasonable considering that Office 365 is still selling Azure compute power - just in a preconfigured setup.

[roselan]

Re: Is Office 365 a separate product from "Azure"

extract from their Q4 2014 Earnings Release.

Commercial revenue increased $1.28 billion or 11%, driven by growth in both our Commercial Licensing businesses and Commercial cloud services. Server products revenue, including Microsoft Azure, grew 16%, and Office Commercial revenue, including Office 365, grew 4%. Commercial gross margin increased $984 million or 10%.

Commercial Licensing

revenue: 11,222

gross margin: 10,296

[Anonymous Coward]

Re: Is Office 365 a separate product from "Azure"

Thanks, too lazy to read the press release.

[roselan]

Embrace, extend and extinguish

Still might be MSFT moto.

Oddly, trusting data to amazon feels like the safest bet. MSFT history of nixing customers makes it a no go. Google is that giant sticky octopus. Amazon only sell books and want to make a penny with it's "superior" infra. Only Morricone theme is missing now.

Next cloud entrant might be facebook, for the lols.

That said, we see two types of cloud: "Product" clouds (oracle, force, adobe, 365), and "generic" clouds. MS cloud growth might come for 365, as there is license and license. Nobody without internal info can tell thou.

[Anonymous Coward]

Re: Embrace, extend and extinguish

"Google is that giant sticky octopus"

Yes, lol @ the recented quoted on The Reg "Ethics? Search Us!" as The Borg's new motto.

[Anonymous Coward]

Re: Embrace, extend and extinguish

"trusting data to amazon feels like the safest bet. MSFT history of nixing customers makes it a no go"

What history of nixing customers? If you look at the long term, Amazon have a significantly worse stability record than Microsoft with a number of major extended and multi-region outages.

[Semaj]

Re: Embrace, extend and extinguish

"What history of nixing customers"

Azure Federations, Azure ACS. Been hit by both of these.

[Anonymous Coward]

Re: Embrace, extend and extinguish

"Been hit by both of these."

I assume you mean you have been hit by service outages. Yes, shit happens. Microsoft have had a few major ones, but so did Amazon in their earlier days. I don't think that risk is significantly different between major providers these days.

[Semaj]

Re: Embrace, extend and extinguish

"Been hit by both of these."

"I assume you mean you have been hit by service outages"

No, service outages happen, that's annoying but it happens.

The 2 services I listed were extolled as being best practice for a while before the announcement that they are being deprecated. ACS will stay around but will not be developed so it's going to just get more and more out of date and Federations are being actively canned next year (not sure of the details so existing ones might still work but I doubt if you'll be able to make new shards).

[Anonymous Coward]

Re: Embrace, extend and extinguish

"The 2 services I listed were extolled as being best practice for a while before the announcement that they are being deprecated. ACS will stay around but will not be developed so it's going to just get more and more out of date and Federations are being actively canned next year (not sure of the details so existing ones might still work but I doubt if you'll be able to make new shards)."

ACS is transitioning to Azure Active Driectory: http://blogs.technet.com/b/ad/archive/2013/06/22/azure-active-directory-is-the-future-of-acs.aspx

The way you do Sharding on Azure is changing - feature updates happen - see - http://azure.microsoft.com/blog/2013/09/05/dal-sharding-of-rdbms/

[roselan]

Re: Embrace, extend and extinguish

I was more thinking of price hikes, visits for MS License lawyers, discontinuing support (not only xp...)

amazon might be the ugly, but MS is definitely the bad.

[Anonymous Coward]

Re: Embrace, extend and extinguish

"I was more thinking of price hikes, visits for MS License lawyers"

You must be confusing Microsoft with Oracle. Microsoft are more often than not amongst the cheapest / lowest TCO options versus other vendor supported alternatives.

"discontinuing support (not only xp...)"

Microsoft provide longer support cycles than the vast majority of other vendors - and the minimum support commitments are well documented. XP was supported for over 13 years! And paid support is still available.

[Trevor_Pott]

Re: Embrace, extend and extinguish

Amazon may have had a few outages - and frankly, Microsoft's uptime is worse - but outages are not related to nixing (or PlaysForSureing) customers. Microsoft is the one with the reputation for overt condescension towards customers, developers, partners and staff combined with a history of cutting every valuable program and reneging on ever important promise they've made.

Amazon has not - to my knowledge - screwed their customers in grandiose Redmondian fashion even once. Microsoft are not trustworthy. Amazong might be. And "might" is a hell of a lot better than you'll get from Microsoft. Microsoft will betray you, me and everyone else. It's only a matter of time.

Microsoft can not be trusted. And they don't give a fuck about that, either. They have no intention of earning anyone's trust because they don't feel they need to do so.

Cloud vendors? Anyone But Microsoft, thanks. I've learned my lessons the hard way.

[CadentOrange]

Re: Embrace, extend and extinguish

Microsoft will betray you, me and everyone else. It's only a matter of time.

I think Google is worse. I've stayed clear of Google App Engine specifically because of the trail of dead projects left behind by Google. Doesn't excuse Microsoft though.

[WP7Mango]

Re: Embrace, extend and extinguish

"Cloud vendors? Anyone But Microsoft, thanks. I've learned my lessons the hard way."

Typical ABMer response. That fact is that Microsoft is competing in the cloud market fair and square. It's called competition, and Microsoft happens to be doing very well, much to the annoyance of the ABM crowd.

[Trevor_Pott]

Re: Embrace, extend and extinguish

I used to be one of Microsoft's loudest and proudest proponents. It should make you stop and think "what could have happened over the years that made him so vehemently anti-Microsoft, especially when he readily admits Microsoft has great technology?"

Or rather, it would make most people stop and thing. I suspect that you are not willing to engage critical thinking with regards to Microsoft.

[Anonymous Coward]

Re: Embrace, extend and extinguish

"Microsoft can not be trusted."

Microsoft actually make money in this space. Guess who is likely to be around longer?

[Trevor_Pott]

Re: Embrace, extend and extinguish

What the fuck does "making a lot of money" have to do with being trusted? The mafia makes a lot of money too, because they will break your fucking kneecaps if you don't pay up.

[CAPS LOCK]

Stifel Nicholaus?

Here's what the Securities and Exchange Commission had to say about them: "and the firm hid the risks “through a series of falsehoods and misrepresentations." "

[Anonymous Coward]

Re: Stifel Nicholaus?

Pretty standard for a US based financial business then. Actually what the SEC said was a bit worse than that:

"In effect, Stifel and Noack persuaded the School Districts essentially to insure the performance of a select group of corporate bonds, and do so with public funds.

Stifel and Noack induced the School Districts to invest in complex financial instruments through a series of falsehoods and misrepresentations. Stifel and Noack knew that the School Districts were risk-averse, and they knew that the preservation of capital was of paramount importance. They materially misled the School Districts about the risks of the investments, the likelihood of defaults, and the safety of their principal.

Stifel and Noack made sweeping assurances to the School Districts, representing that it would take “15 Enrons” – meaning a catastrophic, overnight collapse of 15 investment-grade companies – for the investments to fail. They represented that 30 of the 105 companies in the portfolio would have to go bankrupt, and that 100 of the top 800 companies in the world would have to go under, before the School Districts would suffer a loss of their principal. They represented that there would be breadlines, and that the country would have to suffer an economic collapse greater than the Great Depression, before the School Districts would lose their money. They represented that the investments were safe and similar to U.S. Treasuries.

All of those representations were untrue. Stifel and Noack materially misrepresented what it would take for the School Districts to suffer a complete loss. In reality, Stifel and Noack put the School Districts in jeopardy of losing their entire investments, based on much lower default rates than Stifel and Noack had represented.

Stifel and Noack also failed to disclose material facts. They did not disclose that certain CDO providers had expressed concerns about the investments – including concerns about heavy use of leverage, and concerns about selling this product to this type of investor – and declined to participate for those reasons. They also did not disclose that the portfolio in the first transaction performed poorly from the outset, with a number of the credits suffering downgrades within weeks of closing."

[Nick Ryan]

Re: Stifel Nicholaus?

Colour me confused, what exactly does "Stifel Nicholaus" have to do with a comparison of speculated revenues for MS Azure vs Amazon cloud?

[CAPS LOCK]

Re: Stifel Nicholaus?

A Google search soon finds relationships between MicroSoft and Stifel. For instance: http://www.stifel.com/investor-relations/board-of-directors/michael-w-brown

I think that tells you all you need to know about Stifel.

[Anonymous Coward]

Re: Stifel Nicholaus?

"Colour me confused, what exactly does "Stifel Nicholaus" have to do with a comparison of speculated revenues for MS Azure vs Amazon cloud?"

"Azure grew its commercial cloud services revenue 147 per cent on an annual basis in the second quarter of 2014, and has an annualised run rate of $4.4bn. So says financial analyst Aaron Rakers, the MD of beancounter haus Stifel Nicolaus."

[Anonymous Coward]

Competition abounds

This isn't just a two cloud race. Amazon has competition from IBM, Rackspace, HP, the phone companies, cheap hosters, private clouds and others. The adoption rate of cloud users is also a factor. With more choices and a fairly steady adoption rate, it is going to be a hard road ahead for AWS and they know it. Their UI is way too complex and it makes it look like a hodgepodge of services, rather than an integrated set of solutions. If they continue to drop prices and race to the bottom, well, that certainly isn't sustainable, now is it? And, they have no other profitable revenue streams to support them through price wars that drive the weaklings out of the industry, so there's that little tidbit. The big question around Amazon is; how long can an unprofitable company continue to exist?

[Anonymous Coward]

Amazon is plenty profitable Re: Competition abounds

Why do people keep saying that Amazon is unprofitable or barely profitable? It generates billions of dollars a year in free cash flow; it's more profitable than Walmart, Costco, or Target.

And then it chooses to reinvest that free cash into infrastructure -- distribution warehouses and data centers -- so that it can make even more money next year. How much nominal reported-for-tax-purposes profit it makes is largely a matter of choice at this point.

So really, the big question around Amazon is: why do so many people who can't read a balance sheet have an opinion about its financial health?

[Steve Davies 3]

Re: Amazon is plenty profitable Competition abounds

This --> http://www.reuters.com/article/2014/07/26/us-amazon-com-ambitions-analysis-idUSKBN0FU2DQ20140726

might enlighten you as to the true state of finances at Amazon.

Sure they have oodles of $$$ in cashflow but they are presenting very little in the way of profits to their inversors.

Personally, I think they are getting rather too big for their own boots especially if they are planning on starting their own UK delivery service. That is a whole can of worms that I'd really not want to open.

[Anonymous Coward]

Re: Amazon is plenty profitable Competition abounds

Colour me clueless but are you suggesting Amazon represents a new paradigm that we might all pursue as self-interested individuals?

You suggest that technicalities such as tax ( to Hector or otherwise ) or admitting to making much more than zero are a matter of personal choice?

That I might in fact elect to operate cash flow negative or thereabouts so that I might get that robotic mower I am hankering for?

That until such time as the idea of handing money over to someone other than myself becomes appealing , I can continue in this manner largely forever in order that I might pursue a more pleasing existence generally?

What's the catch?

[robin thakur 1]

Re: Amazon is plenty profitable Competition abounds

Tax laws (generally) allow you tax breaks if your company reinvest in "Research", "Plants and Materials" and other costs, expenses that gets deducted before Corporation Tax is levied. Therefore on revenues of $21 billion it is actually in their interests to minimise their profit, because otherwise, heaven forbid, they might have to consider distributing some of it back to Shareholders...or worse, paying tax on it!

[Anonymous Coward]

Re: Competition abounds

They don't need to profit, just earn enough revenue to pay for data center growth. Amazon has the whole of various industries supporting and investing in their IT capacity that they really only need for five weeks out of the year. The Christmas shopping season. Other firms have to really worry about profitability year round.

[Anonymous Coward]

Re: Competition abounds

"This isn't just a two cloud race."

In terms of large scale volume it is. And in terms of the Gartner Magic Quadrant.

[Steven Marsh]

Google will overtake Azure

Google is the long term winner, they have the best infrastructure.

Sure AWS and Azure are bigger, for now, but that is what we should all want, Competition!

Amazon has the problem of underpaying their people and having a weird company culture.

[Anonymous Coward]

Re: Google will overtake Azure

"Google is the long term winner, they have the best infrastructure.

Erm, in what regards is it best? I can't think of any aspect of Google's infrastructure that is 'best'. They have a far more limited stack than Microsoft and Amazon.

Google are failing miserably in this space at the moment and Ican't see that changing any time soon.

[Wibble]

Loadsa AC's in here...

Strange to see so many people posting AC, especially when extolling one particular viewpoint.

Nice to see Trevor feet-first in the mêlée!

It's good to see some people with experience of using Azure counterbalancing the AC's.

Are there a lot of people posting from Reading?

[WP7Mango]

Re: Loadsa AC's in here...

Microsoft is competing fair and square in the cloud market, and it's clearly doing very well. No need for 'softies to post here in any kind of defence, because the product speaks for itself. Azure is doing well because it's a very good product and offers good value for money.

[Trevor_Pott]

Re: Loadsa AC's in here...

Azure absolutely is a very good product.

It is not good value for money. "Cloud computing" in general isn't. Unless, of course, you're comparing it only to managed implementations of the most high-margin enterprise gear. Cisco on Cisco with a side of EMC and a thick layer of Oracle. Then they're about the same.

But hey, if you want to massively overpay for your IT, that's up to you. Cloud or no, you have opportunity to do so.

[Anonymous Coward]

Re: Loadsa AC's in here...

"It is not good value for money. "Cloud computing" in general isn't."

We can see why you are not in IT management. All those companies must be moving ooodles of stuff to the cloud because it's more expensive. Right?

[Trevor_Pott]

Re: Loadsa AC's in here...

For the most part...yes!

Companies move things to the cloud right now in large part because the monthly billing mechanism allows them to bypass internal purchasing regulations. It's politics - not prudence - that makes cloud computing attractive.

For the very, very few that see cost benefits from the cloud, you have to see what they're comparing it against. VCE versus cloud comes out roughly a wash. Running your own Supermicro + VMware setup absolutely and emphatically isn't going to be cheaper.

Christ, man, it costs over $1500 a year to run a single 2GB Linux VM in the fucking cloud! That may be cheaper for a single VM than buying a server, setting up a place to run it, dealing with backups, DR, etc. But as soon as you get to 5 VMs that isn't really the case any more. By 10 VMs you could be running a completely redundant setup that needs intervention only a few times a year and can run hundreds of "2GB Linux VMs".

So no, cloud computing isn't cheaper. It's easier...for managers. Which makes it politically expedient, nothing more. It's outsourcing by another name, and the people claiming "cloud computing is the solution to all ills for all companies of all sizes" are no different in any way than the shysters who trumpeted outsourcing to India as the solution to all ills.

They were wrong, and so are you.

Now, I at least put my name to these comments and predictions. Why should anyone believe you when you don't have the conviction in your words to attach your name to it? Your assertions and tautologies nothing. As do you, until you can stand up and say who you are and why you assert what you do.

[Anonymous Coward]

Re: Loadsa AC's in here...

Round and round we go.... Every iteration is new to the recently initiated in IT, but is old hat to the old hands once we step back a bit. (A shade of amanfrommars there.) Damn I feel old every time we loop back again be it hardware or software or some fusion of the two.

[Trevor_Pott]

Re: Loadsa AC's in here...

Yup. And eventually, cloud computing might come down enough in price that it's worthwhile for any but the largest enterprises. (Or those needing to bypass politics.) Eventually, we'll even see clouds with zero American legal attack surface, so the rest of the world can participate.

But eventually isn't now. Jumping up and down to cheerlead "this is the panacea for all ills" is ridiculous. Needs are diverse. Risk aversion is diverse. Capital availability is diverse, and the density of red tape is variable. Mainframes are still around for damned good reasons and clouds will be additive to existing IT schemas, not something that will block supplant them.

"It depends" isn't something that zealots can cope with. It must not only be black or white with them, it must be their preferred ideology.

If and when the cloud is ready for the mass market, I'm sure we'll see wider adoption. But right now, today, it's a terrible plan for most businesses if you care about value for dollar, let alone data sovereignty!

[Anonymous Coward]

Re: Loadsa AC's in here...

"Companies move things to the cloud right now in large part because the monthly billing mechanism allows them to bypass internal purchasing regulations. It's politics - not prudence - that makes cloud computing attractive."

You clearly have no knowledge of what you speak. Public cloud services are almost always vastly cheaper than the fully loaded costs of building, hosting, licensing and supporting infrastructure and applications in house - unless you are large enough to have massive economies of scale and are happy to run all of your support and technogy out of third world countries...

For instance it is currently cheaper to subscribe to Office 365 (that includes the full Office client!) than to license Office on it's own - and that's without even looking at the infrastructure, license and support savings!

"VCE versus cloud comes out roughly a wash. "

Lol - VCE is one of the most expensive platform capex. choices that you can make - you pay for fast deployment and preconfigured standardised tested systems and are forced to use the most expensive Hypervisor. The cost per VM is usually going to be strastosherically higher than public cloud. You clearly have never been anywhere near a senior role that has the information, knowledge and experience to make these type of decisions.

"Christ, man, it costs over $1500 a year to run a single 2GB Linux VM in the fucking cloud!"

On which planet? I pay £23.68 per Linux VM a month before discounts:

https://azure.microsoft.com/en-us/pricing/calculator/?scenario=virtual-machines

[Trevor_Pott]

Re: Loadsa AC's in here...

" Public cloud services are almost always vastly cheaper than the fully loaded costs of building, hosting, licensing and supporting infrastructure and applications in house - unless you are large enough to have massive economies of scale and are happy to run all of your support and technogy out of third world countries..."

Prove it you fucking shill. You assert, you assert, you assert. But actually prove it's cheaper for more than just the enterprise. And not using bullshit inflated numbers for OpEx. Prove it or go drink a gallon of fucking draino.

"For instance it is currently cheaper to subscribe to Office 365 (that includes the full Office client!) than to license Office on it's own - and that's without even looking at the infrastructure, license and support savings!"

Again, only if you buy into Microsoft's bullshit "upgrade every three years" fuckery and you honestly think there's value in having a new UI foisted upon you every fucking upgrade generation where you have no say and you - at best - can hold onto one version back. Unless you're a large enterprise, your upgrade horizon is 5 years, and for many companies it's 10 years.

"On which planet? I pay £23.68 per Linux VM a month before discounts:"

Okay, so the prices have come down some in the intervening few months. To get 2 GB of RAM - not 1.75 - I have to choose a 3.5GB instance. That means $892.8 per year, which is still fucking outrageous. That's $4464 for a single fucking 2GB Linux VM over a standard 5 year refresh cycle.

And, in addition, I do actually have to worry about backing up my VMs and data, lock-in and more. So my costs are higher than that $4464 per year, often up to double.

It is you, sir, who has clearly "clearly have never been anywhere near a senior role that has the information, knowledge and experience to make these type of decisions." Which makes sense, for a paid marketing coward who is afraid to put his name to his blithering idiocy.

Repeatedly asserting Microsoft's marking bullshit doesn't make it accurate.

[Anonymous Coward]

Re: Loadsa AC's in here...

"Prove it you fucking shill. You assert, you assert, you assert. But actually prove it's cheaper for more than just the enterprise. And not using bullshit inflated numbers for OpEx. Prove it or go drink a gallon of fucking draino."

Says the man making wild claims about which he clearly has no clue. Perhaps you should back your case before asking others to - Office 365 and Azure Linux VMs have already both been quoted as examples proving you wrong - and from you - nothing but rubbish numbers.

"Again, only if you buy into Microsoft's bullshit "upgrade every three years" fuckery and you honestly think there's value in having a new UI foisted upon you every fucking upgrade generation where you have no say and you - at best - can hold onto one version back. Unless you're a large enterprise, your upgrade horizon is 5 years, and for many companies it's 10 years."

No - just versus buying Office outright - and only renewing when the product was no longer supported - it would usually make more sense to subscribe to Office 365. And that's without even considering the infrastructure and opex costs of hosting / managing the associated servers systems that Office 365 replaces. And one of the core advantages of the Office 365 SaaS platform is that upgrades are automatic, regular and seemless.

"Okay, so the prices have come down some in the intervening few months. To get 2 GB of RAM - not 1.75 - I have to choose a 3.5GB instance. That means $892.8 per year, which is still fucking outrageous. That's $4464 for a single fucking 2GB Linux VM over a standard 5 year refresh cycle."

So you were wrong and now you are just being a pedant to boot. 1.75GB would work for 95% of applications where you might have quoted '2GB'. If you want to price the real equivalent of 2GB then pro rata it = £27.06 per month. The point is that it would cost the vast majority of companies way more than that to host the same VM internally.

"And, in addition, I do actually have to worry about backing up my VMs and data, lock-in and more. So my costs are higher than that $4464 per year, often up to double."

Well you will be glad that backup and disk storage is another great example of where the cloud is almost always cheaper than doing it yourself. you can backup your data from £0.18 per GB per month with Azure (on line storage) down to $0.011 GB per month with AWS Glacier (offline storage).

"It is you, sir, who has clearly "clearly have never been anywhere near a senior role that has the information, knowledge and experience to make these type of decisions." Which makes sense, for a paid marketing coward who is afraid to put his name to his blithering idiocy."

I have told roughly you what I do - and specifically that I have nothing to do with Microsoft. If you don't believe me then that's entirely your problem.