Key Questions – Disruptive Innovation in IAAS

Caution: Makes assumption that you are familiar with few theories associated with Innovators Dilemma.

So, this is just for fun and would love to hear your thoughts on each key question below and if my assessment is correct.

I did a quick 2 minute read of ‘Management Discussion’ of quarterly results for VMW and RAX. I did not do the same for AWS, as Amazon does not pay enough attention to AWS in their quarterly discussion, so I had based it on my impression of them.

Here are the key questions to ask when determining who has an advantage with disruptive innovation – can the incumbent co-opt the innovation or does the disrupter has the advantage.

Key Questions AWS VMW RAX
What are industry players’ business models? Low margin High Margin High Margin
What are their motivations? sell infrastructure to broad base Sell ELAs sell infrastructure to broad base of customers
What are their skills? Build infrastructure for retail Software Build server infrastructure
How do industry players compare to one another? Seen as cloud leader Seen as Virtualization leader Seen as  leaders of Hybrid computing
How do they compare to the needs of the market? Where are there symmetries? Appeals to dev that wants to avoid IT control, smaller shops, Departments looking for efficiencies. Appeals to large IT to optimize their physical server resources using virtualization software. Appeals to IT that wants to  host their servers – does not have dev experience
Where are there asymmetries? Does not sell software or does not sell physical infrastructure Primarily sells software – lacks experience selling service Primarily sells space and IT services – does have experience selling services
Do the asymmetries tilt in favor of the attacker or the incumbent? Favors AWS Disadvantaged Unclear
Does the innovation naturally fit its target market? Yes  No ( vCHS is cramming )  No ( OpenStack is not comparative to AWS)
Is there evidence of cramming? No Yes Yes
Are there signs that a company is ceding a low-end market and trying to move up? No Yes No
Is there an “up” to move to?  Yes ( build custom private clouds ) Yes ( ELAs ) Yes ( Enterprises )
For how long?  ?? Next 10 years? Next 10 years?

Based on this, I think VMW is ripe for disruption, their vCHS effort is Cramming. AWS vs. VMW is an asymmetric fight that favors AWS.

RAX from a strategic perspective does not look too bad, but they need seem to be in need of better execution. AWS vs. RAX does not appear to be too asymmetric, but I am thinking I am missing a thing or two here.

Post your thoughts in comments.


Reading Tea Leaves

During periods of disruptive technology growth, it becomes harder to read tea leaves. In fact, this is the period when incumbents find more convincing evidence that the disruptive technology is just a fad and while it may take off in a decade, it does not pose immediate threat. For ex: if you are an executive selling on-premises hardware and/or software, you look at the following signals:

1. AWS is willing to build private cloud for CIA. AWS is increasingly adopting a private infrastructure friendly attitude.
2. SFDC has discontinued its ‘No Software’ logo and is now starting to endorse on-premises software and also announced dedicated infrastructure based on HP hardware ( SuperPod )
3. Microsoft made several enhancements in both Windows 2012 R2 and Azure to make interoperability between on-premises infrastructure and Azure Cloud
4. Significant investment from many technology vendors into OpenStack and CloudFoundry, which more often than not will end up as an on-premises deployment in a SaaS or enterprise company.

So, you look at these signals and would think these are speed breakers to public cloud and the reality is that there is still lot of life left in on-premises infrastructure. As you start entertaining the idea of lot of life being left in on-premises solution, you start seeing more data points from your existing customers validating your thinking. This leads you into a self-feeding premise that your strategy of being cautious with Cloud is the right one, and you are the wiser one, not the disruptors. You may even think of shorting few Cloud company stocks.

This my friends is what makes disruption so much shocking to incumbents, they ask how could this be – we had all the right signals that our strategy is the right one. They painfully conclude that the tech buyers are irrational, and sane people can’t win.

While this is happening, there are other signs that are coming at you. But, because you are in this self-feeding premise of being right, you ignore them. For ex: this report  from Barclays that predicts the decrease in IT spending as early as 2014 due to adoption of public Cloud.

This report apparently say:

“Our research indicates that traditional revenue streams in hardware, software and services could face even more pressures in 2014, but (the cloud) still creates opportunities for the next wave of tech companies.”
Incumbents  also ignore other signs, for ex: loss of an enterprise customer to a Cloud vendor as an exception or an unusual blip. They also would be ignorant of the Cloud adoption that is happening elsewhere, because these buyers do not view the incumbent vendor as a Cloud player.

I don’t have any recommendations here for you on how to avoid this for your company – you can try reading Innovator’s Solution, but it has been my experience that following the advice in that book is not for the faint of the heart – it takes a man of steel and majority of execs are not that.

However, this upcoming disruption is not bad – its actually good for our industry – while it will be painful to see few of the traditional IT jobs reduce, it will benefit the overall economy through cost savings and efficiency. Also, it makes for lot of fun – expect 2014 to be filled with earnings warnings from traditional vendors and blame game in full swing at these companies. My guess is they will say its because of lazy salesmen.



Couple of Definitions

These are not industry definitions – just how I view things – I have been known to get things wrong, so don’t be running around preaching these as gospel….

IAAS: Infrastructure as a Service

Litmus test: you do not own the infrastructure. You do not have CAPEX. Its a service that you buy from a vendor and you have a monthly/yearly bill as OPEX.

You do not upgrade IAAS, you do not care what manufacturer’s chip  is in your IAAS server – actually, you don’t care about individual server hardware.

Who qualifies? Not really anyone – not even AWS, after some bozo there decided to offer ‘Intel Inside’ instances. Its like a compiler requiring you to use bit of assembly code to get things done. Remember that C used to support it, and had created horrible platform dependent code issues. Yeah, AWS is making an abstraction mistake and they may pay dearly for this mistake.

May be Savvis does with Tier 3 addition, but I haven’t digged into it more.

PAAS – Platform As a Service

Litmus Test: A platform that developers use to develop, test, deploy, run applications – developers or ops do not manage this PAAS instances – you do not own servers, you do not buy VM software, you do not worry about infrastructure.

Who qualifies: SFDC almost does, but this SuperPod kind of makes them look dumb. Microsoft qualifies, but there are doubts on how much on-prem dependency they have. OpenShift online qualifies, CloudFoundry in a hosted manner qualifies.

Few more for good measure:

Private Cloud (n): A pink colored unicorn that spits rainbows.

SuperPod(n): A specially manufactured medical ventilator manufactured by SFDC to keep HP hardware business stay in business. Its strictly done to keep the patient’s family happy, fully knowing patient is terminal.

Single Tenant Multi-tenancy (v): Process of losing your brain cells because you hangout too much with new-age miracle-ists.

OpenStack and Future

I know I really should catch some sleep after that GreenDay concert, but there is so much chatter going on regarding OpenStack, had to get some of my quick thoughts in. This will be quick. 

OpenStack is not dying on the vine. Neither is it thriving. Its somewhere in between. Its good technology in the core, but some bad choices have been made lately. ( Ignoring @littleIdea’s advice is 99.9999999999999999999% of the times is wrong, they should not have done Ceilometer, Solum is ill advised as well ).  So, is it the beginning of the end for OpenStack? No. But, its definitely at cross roads.  When Alessandro of Gartner says it has issues, you need to stop and pay attention ( I had interacted with Alessandro in a prior role and had learned to respect his opinion – for those who don’t know, he was the man behind before his Gartner days – smart guy ).

So, what should OpenStack do? Three things:

1. Give up the idea you are fighting Amazon – you are not, at this point in time, you are like a baby learning to speak and wants to take on Sumo fighter – we get it you have ambition, but, lets be real, people generally laugh at you when you say you compete with AWS. Instead, figure out how you fit into the Cloud ecosystem, which includes AWS as one of the large entities. Do the AWS Interop that people have been arguing for. 

2. Focus, Focus, Focus – Cut the crap down – interview enterprises, find out what is missing, find out what is problematic – fix it. Go into a fix mode, aka Obama’s TechSurge.  Do it now and Q1 ’14.

3. Focus on Community – community is not only about career enhancing acts – its also about caring about the project itself – work on OpenStack core – figure out how to keep vendors at an arms length for next 6 months ( see pt #2 ).

If you do this, you got a chance. If not, Its going to be the ‘what it could have been’ forever. And trust me, failed projects does not enhance anyone’s career and not help companies bottom lines either. 

Ofcourse, you don’t have to listen to me, but I am that little guy cheering on you guys to succeed, so we have a good alternative to a single vendor IAAS. A world of monopoly is really bad for Innovation and I want to see you guys succeed. 

AWS vs. Rest

Remarks by village drunk prior to a village cock fighting contest.  ( language modified by Google Translate: Drunk -> English )

As you guys know, I am a big fan of good theory. In the management science of today, the best insights are offered by Clayton Christensen – if you haven’t read his books, do yourself a favor and read them now.

In this talk, I want to leverage two theories to explain a helpful way to look at AWS.

Theory 1: Disruptive Innovation Theory

Take a look at this diagram that illustrates the theory:

So, the mistake several of us make ( including myself initially ) was to treat AWS as a low-end disruption, meant for non mission critical mission loads. I believe its actually a new-market disruption.  Why?

Per Clayton Christensen,

New-market disruptive innovations, can occur when characteristics of existing products limit the number of potential consumers or force consumption to take place in inconvenient, centralized settings.

Previous state of the art was centralized IT setting up servers and/or virtualized infrastructure for app developers to use server and storage resources. This was both inconvenient and is centralized. Remember the pain of VMware Lab Manager?.

Secondly, parse the words of AWS executives. When they talk, they always use language such as “You could not do this before easily – we are enabling a whole set of new innovations for people that was simply not possible before”.

Is AWS a low-end disruption? No – it has been proven by many vendors that AWS is often not the cheapest IAAS option. There are many tables on the web that shows cost comparison. AWS dropping prices is not related to their attempts to being the low-cost provider – it is passing on their savings to their customers.

So, you with me so far?

Theory 2: The Resources, Processes, and Values Theory

This is the second important theory. Take a look at what this says here:

The RPV Theory

This theory is important from a competitive perspective. Does IBM, RackSpace and VMware have problem with Resources? The answer is a resounding No.

They have talented people, good technology, decent products, cash and distribution channels.

Do they have good processes? The answer is bit unclear – their product development processes can use some change to match the speed of AWS development. However, they do have great processes around Market Research and in resource allocation.

They struggle with the Values. When an exec at IBM, RackSpace of VMware looks at Cost structure, they think like typical software vendor. When they look at Income statement, they want to see million dollar contracts, not $5 / day contracts. When they look at customer demand, they do not see the demand for what AWS is offering. For examples, its unlikely they heard a demand for real-time processing of streaming data. Also, the market opportunity when looked from the existing lenses of these companies does not look very attractive.

The disconnect in values for IBM, VMware and Rackspace for the new market being created by Amazon is one reason they struggle to catch up.

The third relevant theory is Value Chain Evolution Theory. I won’t delve too much into this at this time, but it is sufficient to say that the integrated architecture of AWS is a current advantage for them. They control the entire product architecture. This is not an insurmountable obstacle for IBM, but I am just stating that the current advantage lies with AWS.

Hope you found this talk useful – express your opinion in the comments below.


CloudBeers Nov 19

It turns out some of the luminaries of the Cloud will descend on San Francisco next week. Some for Dreamforce, some for other business. But, these are the folks we can learn lot from. These are guys who are in the trenches building new systems or promoting new thinking when it comes to Cloud. 

This wasn’t an opportunity to be missed, so think its a good idea to have #CloudBeers in SF on Tuesday, Nov 19 at 7 PM. I am thinking 21st Amendment would be a good spot, but open to ideas.  

So, if you can make it, please comment below. If you have suggestions for a different location or different time, please comment below. 

If you are a company that want to sponsor few beers, let us know by commenting below. 

Opinion: Why I think AWS Workspaces is a show of weakness by AWS ?

Notes taken by an American Intern at the emergency meeting of Kazhak Village Elders…

Summary: AWS Workspaces is reflection of a weakness in AWS business and it is a great time for competitors to step on the gas in product execution.

AWS Workspaces starts on wrong premise that enterprises want to use VDI, but can’t because its expensive and complex.

The reason enterprises are not doing VDI is not because its complex, its because its not such a good idea.

The idea that you need to have same configuration across all use devices is an idea that should have died back in the 80s. Unfortunately, some of the control freaks of the 80s are still working in IT an while they are becoming a rare breed of IT Mgrs that no one likes, the reality is that most LOB managers recognize higher productivity that can be had by enabling end users and not by controlling end users. This in addition to the employees mixing their consumer apps and expecting similar experience from their enterprise apps ( what some term as COIT) is leading to variety of devices and configurations per user.

Lets say if IT dictates user must use IE8 and user users some personal apps while at work. She may use a third party password manager like LastPass or 1Password. Now, if IT were to do a configuration push every night and reset the configuration and thus lose all the local storage, you would have a pretty pissed off employee in the morning.

Lets say the user has purchased a copy of Balsamiq, because they needed to do quick prototyping, and next morning, IT pushes a configuration refresh – you have a pretty pissed off employee in the morning.

My point is the desktop, laptop is personalized per employee – this is both required and encouraged.

But, you say, I can’t meet my compliance and my auditor won’t be happy. Fuck the auditor. Seriously. If your company goes out of business because you are 100% compliant but not building new products because you have treated your employees like rats, your auditor won’t pay your salary.

So, we looked at it from end user perspective, what about the cost reduction aspect?

Its not like AWS Workspace is magical and works in Ether. You still need to buy desktops and laptops and the laptop suppliers have deals with OS manufacturers like Microsoft to bundle the OS with the machine. So, you pay for the machine  and then you have to pay like $500 / yr/ user ? Thats not very smart, is it? You can use your own internal calculators to arrive at your cost savings, but my guess is, you won’t be saving money.

Now, lets look at each use case AWS Workspaces promises to solve:

1. Mobile Device access

So, when has it become acceptable to force a Windows 7 experience onto an iPad?

2. Secure Workspace ( compliance )

Addresses the egos of IT control freaks. You need to fire them, not enable them as a CIO, if you want to keep your department relevant.

3. Remote Employees

So, your auditor is going to be happy that an employee is reviewing a patient information on a public train? Really? And a public Starbucks Wi-fi is going to be more secure than data on a laptop? Give me a break.

4. Seasonal Workers

Ok, this is half decent – but, don’t you already have device wipe policies – do you want to pay $500/user/yr for a wipe functionality?

5. Student Workspaces

Ok, this makes sense – but again, this is a problem that has been solved using less expensive and may be less elegant solutions. In this case, AWS Workspace may be the Vitamin you may not be able to afford.

6. Developer Workspaces

Don’t even get me started – you try to control developer environments and force them to use NetBeans, you aren’t going to have much luck hiring programmers, would you?

In fact, when I went thru the use cases list, I kinda felt sorry for the product marketing guy who wrote it – you were trying to tell a story that doesn’t exist. Feel your pain bro. This reads pure desperation – they had to write about a product even they don’t believe in.

So, you ask Why AWS is doing it – isn’t AWS so smart, they can’t ever make a mistake?. I can only think of two reasons and I don’t know if they are right or not.

Reason #1: This is a pet project of Andy Jassy – he wanted a  comparison against his HBS classmate who runs the End User Computing for VMware.  ok, I get it, Business Schools, Big Egos, whats the big deal wasting few millions here and there right. Yeah, I get it. Ivy League thingy. Its unlikely though, since Jeff Bezos is known as a penny pincher, he may not let a thing like this happen at Amazon.

Reason #2: AWS is hitting a ceiling on their growth – they are trying all kinds of stuff to see what sticks. This may be the Google Wave of AWS – they will try this and finding no one is using it, will kill it next year. If this is true, then AWS peak is over and they are on a downward slope and this gives an opportunity for a right competitor to attack them. ( Protip: you won’t do it using a on-premises business model). This is more likely reason and this is an important signal. This is time for would be competitors to step on the gas in terms of product execution.  This also is a signal that AWS does not have a long term product roadmap they are executing towards, but are out of ideas.

Good luck guys.

Note: As someone pointed out, this may be good for some select industries such as call centers, manufacturing floor kiosks. Not discounting that, but those guys are generally last to move towards something like this.  ( Welcome folks to correct me, but from what I see, they are often laggards , but I have limited experience ).

So you think you can compete with AWS?

Short post inspired by Peat infused Scotch….

Often we see different offerings come up and say they are going to compete and be better than AWS. They focus on lot of things, both right and wrong. I want to focus on just few criteria that every vendor who wants to compete with AWS should ask themselves:

1. Can a developer create a new instance using less than 5 clicks without their IT doing prep work?
2. Can an IT guy pick an additional service that is good enough or better and without having to scratch his head about pros and cons of various options?.
3. Can an organization scale their app without worrying about underlying hardware?
4. Can an organization get rid of the servers they don’t need and see immediate cost savings?
5. Can an IT guy buy services without ever talking to a sales man and just using their credit card?

If the answer is No to any of these, you don’t have an offering thats worthy of competing with AWS – please get real and don’t be comparing yourself to AWS. Instead pursue your own market and destiny.

At 5,000 Tweets

Read about the Disaster in Philippines:

Donate here:

I want to tell you three things:

1. Who am I?

An average guy working in the software industry with a big ego. No body famous, not a luminary, just an average guy. I don’t work for any of the cloud vendors you may be familiar with.  Remember that guy in school who goes and writes stuff on the whiteboard before students and teachers arrive? Yup, I am that guy.

2. Why am I doing this?

Because I can :). Seriously, for fun, nothing more. No agendas. I started this as a fun habit, and also as a way to calm myself when I see stupidity or just plain bad marketing. Thats it. Nothing more. Try it, anonymity is a good stress relief.

3. Ok, what do I want you to do ?

If you liked my tweets or if you hated them, this is your time to act. We all live and breathe technology and as a result are somewhat blessed with good earnings. And I know most of you are awesome when it comes to donating to Charity.

May I ask you to do it one more time and donate to RedCross to aid with the Typhoon Haiyan Disaster? Please see here:

That’s it. Thanks guys.

UX Designers considered harmful

Your product is an UX disaster. It has pretty screens, but behind those pretty screens is non intuitive experience. Your end users have more WTF with your product than while driving on the roads. Your product is so bad, your marketing uses cartoonish drawings in promotional videos instead of actual screenshots.

You are told to hire an UX rockstar. You find the smelliest person you can find and hire them, because you are led to believe a rockstar must have long hair, bad BO and also be obnoxious. You bring that person in. The UX rockstar comes in and says things must be flat and he makes every button look and be of exact same size. All your dev team says this is greatest thing since slice bread ( they have had it with tweaking CSS for each button anyways, so consistent CSS is  attractive).

Your users are even more unhappy. You convene a council of elders meeting. You learn that the problem is that UX guy you hired is overburdened and thus engineers are still making up stuff on their own. The way to fix this problem is to hire more UX people. You make your UX rockstar a department head and give them a group of 5-10 new UX rockstars.

You start getting back beautiful power points and expensive expense reports as they attend all kinds of design conferences. They also insist you buy some Van Gogh and other expensive art to inspire design aesthetics in the company.

Your users are even more unhappy. 

Story sound familiar bro?. UX Designer role may be the greatest trick ever pulled by the devil and we are all been taken for a ride. Consider this:

1) Apple design is an exception and your rockstar ain’t Ive and you can’t afford him anyways.
2) Best UX products are often designed by the founders of the companies that have great design sense – you are not one of them ( if you were, you wouldn’t be building UX departments)
3) Acceptable, if not great UX is often done by engineers, rarely by an UX team.
4) 94.17% of all UX engineers do not have a clue about what the users want to do and they don’t even bother talking to users.

These are what I had observed – you could argue my observations are wrong and you would be welcome to do so. I only ask you pay for the drinks when arguing and as you already know, I only drink Ardbeg or better. 

So, what do you need to do:

1. Hire software engineers that enjoy talking to customers. Hire engineers that are humble and are interested in listening to others. These are not easy to find, but totally worth it. Protip: They rarely apply to jobs that have descriptions like ‘rockstar’, ‘ninja’
2. Give engineers a free hand to create mockups based on their interviews with customers.
3. Hire engineers who are experts at front end technologies and can prototype ideas in matter of minute, not days.
4. By default decline all requests to attend design conferences – they are nothing but amplifying the bozosity.
5. Ignore everything Apple is doing – its not for you. Its just not for you. You can’t apply Apple lipstick to your Pig – your Pig needs to find its own style.


Disagree? Post in comments below: