Microsoft Azure Pricing Adds to Cloud Confusion
The long awaited Microsoft Azure pricnig model was announced recently. For a while I really thought Microsoft would use its fashionably late arrival to Cloud Computing to build upon the short comings of other vendors and maybe take advantage of the opportunity to establish a leadership position in Cloud Pricing. Alas, yesterday I discovered that Microsoft will instead add to the confusion when it comes to Cloud Pricing. Here is a snippet of what they announced:
Azure pricing is a disappointment on three counts:
- It isn’t true utility or consumption computing pricing because the customer will not end up paying for what they actually use. Compute and Storage hours are a mistake because it presumes the full consumption of compute unit for any part of 1 hour. The customer doesn’t realize it but they are paying for computing resources they aren’t necessarily going to use. This is the hidden margin for players like Microsoft and AWS. While the customer watches the left hand, the right hand is busy taking money right out of the customer’s pocket! This pricing model is the equivalent to the old salami banking scam. Take fractions of cents from a billion accounts and think nobody will notice.
- It will not support the migration from legacy operating models to cloud operating models. I’ve written about this phenomenon before, but it seems like the big cloud players are dramatically out of touch with the average customer. This pricing model means nothing to the customer but a bunch of numbers. How does this pricing model translate to what the average customer doe today in their enterprise?
- There is absolutely zero predictability. And if there is, it sure as heck isn’t simple enough. Can someone tell me what the cloud will cost me for 15 .Net applications I have built before I get an actual invoice from Microsoft? Hold on, let me call my MIT laureates and Actuarial PHd’s to do some predictive billing!
I will say that Microsoft, unlike a lot of other cloud platforms on the market has smartened up to the fact that they need to account for Storage Transactions. Storage transactions can be the silent killer of cloud computing infrastructure. The only problem of course is understanding what that means to my application. How the heck do I know how this translates to a real world operation?
Cleaning Up the Cloud Computing Pricing Mess
Cloud Computing is the next great land rush and it is happening now. All the major technology companies have their offerings. And it seems like everyone is entering the market – even the hosting companies want in on the land rush.
In theory, migration to the Cloud makes business sense; you’re enabling companies to rent computing power that would cost them too much to buy. I won’t bore you with yet another blog post on the ‘what is it’ topic. There is a great synopsis of Cloud Computing published by Mache Creeger and I recommend checking it out. In our model, we’re allowing companies to pool their resources on the supply side of Cloud Computing and leverage a much bigger, better shared infrastructure on the demand side of the equation.
Cloud Computing is about lower costs and greater use of resources. Greater flexibility, more options and overall, more computing power. It’s a shared cost. And it’s based on what you use. Or is it?
One of the areas of Cloud Computing that still needs to be addressed is the issue of pricing. Pricing the Cloud has gone beyond complex and confusing and entered the realm of ridiculous on some levels. We’ve met with countless service providers in the past year and the basic message is clear: Come back when you can give me something that doesn’t need a PhD from MIT to decipher. This message was also pretty clear at last month’s Interop Las Vegas event.
The odd thing is that everyone agrees that Cloud Computing pricing needs to be standardized. Many companies want this to be an industry group that develops standardization. Industry groups and alliances have been throwing this topic around for a long time now – we’ve seen this question come up for more than two years. But why is it that nothing has happened? As a company that has what I would call truly transparent pricing, I’ve been confused about this for a while.
I was recently on a conference call with a potential data center partner when I got insight into what I truly believe is the answer.
Standardized pricing and corresponding tools that allow end user customers to peer into the rack and seriously drill down into the granular cost of the Cloud are simply bad for business.
In fact, the parties on the supply side of Cloud Computing – elastic computing providers, managed hosting companies, platform-as-a-service shops, big iron manufacturers, etc. – don’t have much incentive at all to strive toward pricing transparency and standardization. Why would magic quadrant hosting providers or heavily vested IaaS providers effectively even the playing field by adopting a standard pricing metric when it is their brand that is ultimately buttering their bread today? Is a company like Amazon or Google really going to adopt the same pricing standard as every other company getting into the race? Maybe, but don’t hold your breath in hopes to see them at the front of the line.
I think the work of Cloud standards advocates like Reuven Cohen of Enomoly has been really great for cracking the nut of Cloud interoperability. But it may be a stretch when they dream of Cloud interoperability extending beyond the technical exchange and integration of systems and data. Here is a reality check: All the big Cloud Computing providers in the market are profiting from preventing the very process of commoditization they allegedly support. And even if you aren’t part of that group, pricing is an integral part of the profit picture and thus cannot be decoupled from the discussion. Just because you get together and document some sort of standard or benchmark doesn’t mean you’ve solved the problem for the stakeholder that matters most (the customer). In fact, I think these types of standards groups may only serve to muddy the waters further on the subject because they don’t pay enough attention to the connection with the bottom line for a Cloud operator.
Understanding the profit motivations of the Cloud providers and then dissecting the current modus operandi for pricing exposes a huge gap that I think will shape a big part of cloud development initiatives in the next few years.
Let me give you an example to prove my point:
Cloud Computing service providers seem to believe that they can and should charge for the Cloud on an hourly basis. On the surface that sounds great, because it’s better than paying for a machine for the whole month, isn’t it? But underneath there is a lot more to it. Think about it. If you use a server for one minute of an hour, you’re charged for the whole hour. That’s crazy. One sixtieth of an hour costs you the whole hour? Sure the pricing is reduced, but what are you really getting? I think Allan Leinwand captures broader implications of this silliness quite well when analyzing the state of Cloud pricing. He said, “CPU hours: that’s not something I go buy. I buy a blade server, and the hours are infinite, they’re mine.” Leinwand has a big point and it has a direct impact on the future capability of Cloud Providers to achieve mainstream relevance to the average enterprise. 6fusion’s CEO and co-founder John Cowan analyzes the implications of pricing on the buying community here in a separate post.
And if it were really just as simple as clocking CPU hours and sending out a bill, maybe we could alleviate this pain point in the Cloud and move on. But it doesn’t end there. Invariably, Cloud Vendors have to “tack on” all sorts of ancillary charges and fees to make money. Everything from RAM to storage to bandwidth and even Support get thrown in as separate line items. The pricing becomes convoluted and difficult to predict. It’s a huge mess, but there is no incentive to solve the problem, given there is a lot of money to be made from the confusion.
I have no problem with the supply side making money. After all, that’s what a company is in business to do. What I have a problem with is the lack of transparency or ability to leverage these systems for anything more than just the technical accomplishment of elastic computing (don’t get me wrong, that is a biggie!). When Cloud providers don’t give you proper insight into what you are using, and if you can’t make the mental jump between what you do today (ex, buy more blades) and what the Cloud represents, the advancement of the industry suffers.
Herein lies the gap.
Service Providers must deliver more insight and transparency into the Cloud, not fog the pricing just to earn more margins for a brief time. Customers are far too smart for this to work long term. Ultimately, we believe that in order for the Cloud to succeed, the industry needs to help customers understand their true usage and the true value they are getting before and after they make the decision to use Cloud Computing to run critical IT systems. A granular metering and billing technology that transcends the politics of brand and vertical silos, while satisfying the need to be a ‘profitable’ service provider, will go a long way to helping to clean up the mess that is Cloud pricing today.


Application Profiling: The Key to Customer Cloud Migration
For the past several months the team at 6fusion has been working directly with a select group of customers using a prototype software tool called the UC6 Profiler. The UC6 Profiler is an agent we created that uses our patent-pending algorithm for measuring utility computing consumption. The UC6 Profiler meters live client applications running in their own offices or data centers, recording resource usage as though the applications were all running within 6fusion’s federated cloud infrastructure. The report output paints a clear cost picture, application by application, giving the customer an unprecedented set of data to guide and support their decision to migrate any or all applications to the cloud. We’ve provided an example of the output report here. It’s early stages yet for this, so the info is pretty raw. We’ll ‘gloss it up’ when it goes into production later in the year.
In our experience, the number one customer question about the cloud is “what does it cost.“ Like others in the field, we see the ability to profile consumption and report running costs to be one of the missing links to cloud computing adoption. As you can see in the sample report provided, this customer can identify that the application called “CUSTMAIL01″ would cost the ‘most’ to run
in the cloud. Conversely, the application called “CUSTAAC001″ would cost the least.
Going beyond the customer implications, the UC6 Profiler could also be the missing link for the IT Service Provider community to truly take the reins of the cloud and leverage it to build significant new revenue opportunities. But the implications don’t stop there. Profiling can play a huge role for ISVs looking to plan and price SaaS offerings. I’ll elaborate on this in another blog post. We’ll focus on moving one mountain at a time!
Here are a few other tidbits we can share with you for now:
Stay tuned for more to come regarding the UC6 Profiler in the coming weeks!