Good debate going, so let me throw some fuel onto the fire.

1.  I specifically did call out that the point I was making was NOT inclusive of operating considerations – and that while those should be included in a USER decision – the reality is they are often not.  Deal with it.

2.  The customer area of focus for my point is the COMMERCIAL IT buyer – the traditional folks who pay lots and lots of money to run – albeit sometimes badly – IT shops.  That’s where the big money sits, and that’s where we all seem to refuse to acknowledge the realities of this nascent market.

The actual buyers of public cloud infrastructure services from the commercial world today fall largely into a few categories; small businesses/new businesses, and fringe/rogue individuals/groups frustrated with internal IT service delivery (those developers who go to Amazon instead of internal IT).  New entrepreneurial ventures – especially those second time around folks – are fast adopters of all things cloud.  As they grow and time moves on, they will become the way, but for the next 10 years, they will remain a small slice of the money pie.

Established commercial entities do not represent a signficant population/market – at least for now – for infrastructure services in the cloud (public) – but instead spend all their cloud money on APPLICATION services (SalesForce) or private cloud services.

3.  Application services in the cloud are “persistent” while infrastructure services tend to be “transient”.  The developer using Amazon is renting short term stuff – but the company using SalesForce is in it for the long haul.  My backup service is far more persistent than my need to rent some compiling cycles.

4.  If you are only providing pure infrastructure services you better be big, fast.  I contend that you will find only a very few providers ultimately – the same way that other major industries have consolidated down to those best positioned to deliver – such as what has occurred in the Telco and Cell phone industries.  The huge investment required for global infrastructure -and the ability to transact directly with the buyer – limits the prospects for most entries to make any significant sustainable independent play.  Regional domination is at best what most can hope for, then an acquisition by a bigger player.  History proves this model out time and time again.

Therefore, I contend that from a pure infrastructure services perspective, the most likely successful entities to remain standing will be the major Telco’s.  They have the bandwidth relationships already – thus are the Internet pathways – that lead to and from the cloud infrastructure, and as such have the best possible position to add services onto their clients’ bill.  As they do today, they will buy their equipment and expertise from the lucky few they choose to deal with – but they will most likely own the customer.  I am NOT saying they are better than you at delivering anything – I’m saying they are in a much better position than you are to do so.

5. That means that if you are providing infrastructure services in the cloud, you best start bringing higher level application value to the party or you are inevitably going to be toast unless you are one of the big giants left standing.  EMC Atmos/ATT is a lot like Apple iPhone/ATT.  We bitch about ATT but if you want an iPhone, what are you gonna do? You don’t see a lot of new entrants into the Cell phone provider business these days, do you?  You need to have a sticky reason for customers to want you – and it can’t be “i’m cheaper” per GB.  That is the point I was making.  If we can argue about what’s really cheaper already, what hope do we have?  Basic economics – little guy can’t be cheaper than big guy unless the big guy lets them.  Fact.  Little guy can be better at something – but unless the consumer can put legitimate value upon that something, you aren’t going to get paid long term.

Part of the problem is humanity and emotion.  There is a very big difference between “wanting” something to happen and it happening.  We “know” that a fifty year old manufacturing company will be better served by not dealing with their IT stuff internally and heading to the clouds, and we “want” them to do so – but that just isn’t realistic.  What “should” happen is normally not what “does” happen.  How is Star Office treating you?

Wishful thinking is what the SSP rage in the early 2000′s brought us – an overfunded, overhyped circus that never had any sustainable legitimate business value – because even though it was “better”, better didn’t matter.  Commercial concerns that had existing legacy operations were simply not going to hand over those functions to someone external, no matter how much they hated it.  The reward never surpassed the risk.

Big commercial entities are telling all of us the deal, but we’re not listening.  If a company has no vested existing interest in IT – they will go cloud right away.  After all, who in their right mind would want to go through this hell?  If they have decades of existing IT operational interests, they won’t – at least not to the public cloud.  The key is to take what you are given – take a piece, and add value.  By providing the customer a means to prove the service worthy, without risk, you enable them to accelerate their could offload decisions.  This is exactly why desktop backup has been a winning cloud offering – not because of the infrastructure, but because IT doesn’t deal with desktop backup!

My friend Stephen Foskett thinks I’m evil for failing to point out that cap-ex is only a small part of the cost picture.  Alas, he misses my point – which is this:  no matter what, for existing IT operations, there will be questions when there are significant cost discrepancies.   And, as soon as someone questions the cost structure of doing it internally with people we already pay, on gear we already support, then the service provider is on the defensive – and that is a shitty place to be.  As soon as the conversation slips from the legitimate value provided to the baseline commodity cost discussion, the game is over.

What Steve should realize, is this is NOT the argument.  Instead, the reason his company has customers has little or nothing to do with the fact that they sell capacity.  It is what that capacity is used for – what application can Nirvanix provide that builds value and utility to the assets residing on it.  He himself commented positively when I reported that things such as auditing were critical gates to commercial banks adopting cloud services (above security, which surprised me) – he rightfully noted that the Nirvanix platforms’ value was built on being able to provide audit ready services.  That is value add – application value add.  Capacity isn’t.

Others commented/noted some other things I  found interesting.  People love to argue with me for saying I can buy 1TB for a few hundred dollars – that it “isn’t the same” as buying mission critical gear.  Sometimes that is true – but not always, not anymore.  One person spoke about how with Amazon, they are getting replication to two sites.  Really?  How do you know that?  Why would anyone assume that Amazon is using anything better than my cheapo RAID stuff?  I would contend that they aren’t.  For their model to make money, they can’t use traditional core mission critical stuff.  That’s the bigger truth here – you don’t know what they use.  They aren’t telling you.  They are telling you that you can have X capacity for Y price.  That’s it.  They could be putting your data on wax rolls spun by monkeys for all you know.

We’ve been raised to assume that if it’s cheap, it sucks.  While historically I agree with this, Gmail and Yahoo Mail changed my opinion on this subject.  You can no longer make the assumption that cheap or free is inferior.  But lets face facts, I have way more trouble with my expensive Microsoft stuff than anything else.  You simply can’t make this argument carte blanche anymore.

The cloud “industry” will evolve to this:

A: Big players will provide 90% of the infrastructure services – and it will rapidly commoditize.  Amazon will continue to dominate the “no value” commodity market, and a few big guys can take the “mission critical” business prize (this is what EMC, HP, IBM, etc. are fighting for).  I don’t see any room long-term for “in betweeners”.

B: Bigger players who own the customer relationship (and plumbing connections) will be the true leverage points.  Telco’s are key.

C: Application value is the key success metric for widespread adoption and new value creation – aka $$$$.  Don’t try to be the next Amazon – try to be the next SalesForce.  You can succeed as an independent if you add legit value.  I’d rather license functionality to Verizon or Orange rather than raise a trillion dollars and attempt to out-do them in their markets.

Let’s call Amazon the modern, wiser version of StorageNetworks.  Just as the other 894 wanna-be SSP’s went away, so will those who want to compete with Amazon on the same terms.  What came out of the SSP era?  Backup service providers.  A legitimate value add and market need.  Capacity providers died, but backup service providers remain a vibrant, healthy business.  Amazon’s wisdom is that they aren’t trying to pretend (at least so far) that their value is anything more than it is – they are commoditizing THEMSELVES so others can’t.  Trying to beat the incumbent in a market where the only value is commodity services is LUNACY.  Adding true value on top of that commodity is where the real value (and $$$) lies – for you and your customers.

In the meantime, if all you offer is a commodity, you best make it unquestionably cheap.  We’ll chat again at your next venture…..

Related posts:

1. Reverse Cloud Economics
2. Why the Cloud will Vaporize
3. Scarcity Continued….
4. Cloud Apps That Make Sense – Hint: Backup isn’t One
5. Cloud Control: Lessons Learned From Iron Mountain

Tags: Amazon, Cloud, economics, EMC Atmos, Orange, SalesForce, StorageNetworks, Verizon