Why OpenStack Solum may be misguided

Fireside conversation with village elders while drinking Ardbeg, which is known to increase your intelligence by multi-fold.

Few things first:

1. We in the village love OpenStack – we think it need to succeed to offset AWS monopoly in IAAS.
2. We love that OpenStack is run by passionate and smart engineers. Lot of respect for you guys.
3. We love that it has attracted companies with deeper packets, so its not going to disappear and is the best chance we have for a future with options.

Now, for some education.

Because we not have Breaking Bad or Dexter to keep us busy, for most part, we sit around the fireplace and discuss strategies. Long time ago, we found the Chasm Model to be easy for describing the market evolution. We also think while Clayton Christensen’s Innovators Dilemma is good for lighting fire under CEO asses, but it causes more confusion among people who haven’t taken time to read it fully. Disruption is real and people who ignore it will learn Clayton’s sayings the hard way. However, its not for the masses.

With that, we  made an observation few weeks ago:

1. IAAS has crossed the chasm thanks to AWS.
2. PAAS has not crossed the chasm and burden of proof is on PAAS vendors.

We had not expanded on it, because our village head thought its self-evident. However, the village council had concluded it had been a mistake not to expand on it. This post will try.

When you are in the Early market, you work with Visionaries ( Early Adopters ). An example of this is customers using CloudFoundry. Look at this post to see how visionaries think. Observe the language – ‘Radical Change’. Visionaries look for ways to break with the past and start an entirely new future. They think breaking with the past will give them insurmountable competitive advantage over the old order. When they succeed,  its a sweet victory.  We as an industry have a lot to thank visionaries, for without them, we would still be using punchcards.

After you had crossed the chasm, you enter the Bowling alley and Tornado.  Here you work with Early and Late Pragmatists ( collectively called Early Majority ). These are typically  companies that like to adopt technologies early enough, but not bleeding edge technologies. These folks do not love technology for the sake of technology. They believe in evolution, not revolution.  These guys like to buy from market leaders. Most in this stage want to buy from AWS, who is the clear leader.

Looking at the revenues of AWS ( expected to be $3B this year) and the rate of growth, the village heads have said that it looks like a Tornado and we are past Bowling alley.

To start, OpenStack already has a disadvantage – the best they can hope for in the current cycle of growth is second place. You role is one of a Chimp. You competing against AWS is competing against the whole market. Do not attempt – you fail. You have permission to get some market share, but not dominate the market in the current cycle. Focus on a subsegment of the market. For ex: companies that want elasticity of the cloud but have preference for either private cloud or hosted private cloud. This is the time for you to differentiate yourself from the Gorilla, which is AWS. Good news for you is there are people among you that are pushing to go in that direction. You just need to listen more. If your sole obsession is fighting AWS, you cease to exist when the tornado period is over and main street begins.

You need focus that is specific to the subsegment of the market you are focusing on. You need to be targeting operations staff that want the benefits of Cloud, but are not comfortable with a operator driven public cloud. They are screaming currently that your install and configuration is not as easy as it can be. Fix that first. That is more important than trying to feature match AWS. Again, focus on the subsegment and do the best job you can for that market. As outsiders, our elders think your strategy looks a bit like spray and pray.

This is not the time for you to fragment your offering to try to appeal to both early majority and visionaries. Solum is your attempt to appeal to PAAS audience who for most part are visionaries. The worst thing that can happen here is that one of them gets interested in Solum and asks you to implement myriad of features they need ( which is what visionaries do). The distraction from trying to appeal to visionaries will distract you from trying to appeal to the subsegment of the early majority.

Think about it guys – if you do not succeed and differentiate yourself to be available when the market goes mainstream, the future looks bleak.  Leave the task of crossing PAAS over the Chasm to CloudFoundry and OpenShift folks – they have the money and talent to do a good job – they are not your enemies right now.

We could continue more, but we just ran out of our supplies of Ardbeg and Azamat is not approve more Scotch budget.


How Not to Suck at Technology Marketing

Helpful tips from a village elder in Kazhak as told to a goat-herder who want knowing how valley companies can suck less.

1. Stick with Facts

No matter how tempting it is to make far reaching claims, stick with facts. Especially with developers. Your far reaching claims will hurt you more than help. Its a funny thing, developers can’t detect their own body odor, but they very good at smelling marketing bullshit from miles.

2. Promoted Tweets are lame

Don’t use them, unless you are selling male enhancement drugs and the topic being discussed is related to it. Technology promoted tweets are quite lame. Especially during conferences, don’t try to insert your lame promoted tweet into a conversation. You will get called for it.

3. If you have customer stories, promote them – if not, just say its your hypothesis

Anything before you have an actual customer is at best a hypothesis, so say it so. Say “we think this technology will help following use cases” –  just because you ran into a CIO at Starbucks this morning and he asked you what you working on, does not mean they are evaluating your technology.

4. Resist the urge to attack your competitor

You need to position against a competitor when you compete with them in a deal, not on a public forum. It makes you and your company look like bunch of idiots. If you are truly awesome, have a customer talk about benefits they had from switching to your solution. Not thinking about switching, but have actually switched and have been on your solution for at least 3 months.

5. Beware of buzzword bingo

There are no awards for use of most buzzwords – every time we hear someone say “Software Defined”, we deduct 10 points from the value we assign to the person talking. So, its risky, be careful. Other things to be aware of are “ground breaking”, “revolutionary technology”.

Now if you are ground breaking, take a picture of the ground you broke and show it, may be we believe you. If you are revolutionary, show us which revolution you caused and document the time you spent in a jail because you caused a revolution. If not, you are just being lame.


Opinion: Four Premises we need discussion over beer.

Mah Opinion – as told to Kazhak Public Radio, translate to English by c-cloud Intern. Before you react and smash the beer bottle on your forehead, know that in a battle of opinions, burden of proof is on you to offer solid data to say why mah opinion is no good.

Premise 1: Only possible cloud in the future is operator driven public cloud.

Questionable.  Consider following:

1. Location of infrastructure and data matters more now than before NSA revelations.
2. Early evidence of price gouging by leading public cloud operators ( AWS reacting by offering reserved instances etc, but some “small” companies got burned saw through it and are leaving. We hear people talk in hushed tones that AWS may not represent such a great value).
3. There really isn’t a great operator competition – AWS is a leader by clear margin. Google is mostly engineers running around like chickens with their heads cut off. Microsoft viewing cloud as way to protect their Exchange, SharePoint and Office revenue streams.  No serious competition may emerge in next 5 years.

Premise 2: Private Cloud can not succeed


1. For next 10-15 years, private cloud is going to be the preferred model.
2. For well defined services, SaaS will be adopted – do not confuse that as public cloud adoption.

Premise 3: Any Private Cloud should be anchored to the leading public cloud to survive


1. Not enough compelling evidence that customers worry a lot about this. ( They want it, as our previous survey shows, but its not a show stopper ).

Premise 4: Cloud market is going to consolidate to few public cloud operators

True – but it will take 5-10 more years before that happens.

Ahoy Platforms…

Note: This blog post modified to display properly on your outdated monitor ( get the Retina display already dude). Language modified by the furloughed guys for your reading clarity.

It is said that when a vendor strategy is not clear, they start calling their product a platform. A platform is not always perfect and neither is it stable.

Platform captures that sense of dreaminess, a potential that your tiny brain can’t yet grasp. It is also an  attempt to discover your connection to nature. Also know that the word that follows “platform” is “ecosystem“, which with its lush green leafs, stems and branches and  beautiful green frogs jumping from  ponds, birds singing beautiful songs. And often times when the ecosystem is so green, it is expected there some rainbows in the sky too. For added effect, imagine a pink unicorn freely roaming the ecosystem grounds.


So, in someways ‘platform’ is the romanticism that has been missing in the technology industry. It is also the  attempt of marketers and BD folks to try to bring back love, Kumbaya and all that feel good stuff back into our engineering driven culture.

How is an IT guy who is evaluating products to solve their specific problems avoid getting sucked into this ecosystem of frogs, birds and become giddy eyed?

we give you some guidance, which you shall follow at your own risk…
1. Platforms are temporary –  platform don’t lasts as long as your systems have to
2. Look at multiple platforms and adopt a strategy of heterogeneous platforms
3. Ask to see interoperability between platforms – if nothing else, vendor dance is often fun to watch
4. Do not fall for ‘but, its open source, so lets hug’ trap
5. Use the core value provided by the platform, but avoid the other services
– for ex: do not fall into using the DB provided by this “platform” – you are better off getting it from a vendor that knows databases ( SQL or noSQL )
6. Good idea to avoid a platform for first 12-18 months of its initial existence, unless you like causing yourself pain and somewhat masochistic about it.
7. Listen carefully to expert’s advices – memorize them, their sayings are good to repeat with friends after few drinks when you run out of jokes.
8. Look at few images from rain forests and few other places and create strong association between ‘ecosystem’ and that image – when you hear ‘ecosystem’ the first thing that your mind recalls should be a pretty green frog jumping from a pond.
9. Avoid PPT BS and platform promises and rather insist on a POC that solves your current problem well.
10. Know that many IT managers before you have resisted the lure of platforms and they have done well for themselves – missing a platform wave/quadrant is not that big of a deal.

And to retain the sense of disdain for all things Platforms, you must use the word ‘platform’ at least once on twitter.

IT Trends Survey

Results so far – let your voice be known here: http://www.surveymonkey.com/s/Q8TYTC8

( Will keep the survey running longer )

1. Are you happy with your IT?

Yes – 29%

No – 61.29%

2. Does your IT Embrace latest technologies

Yes – 29%

No – 64%

3. Do you think changing CIO title to something else will fix all the problems?

Yes – 29%

No – 64.5%

4. Do you think IT department will cease to exist in:

5 yrs – 12.9%

10 yrs – 10.9%

20 yrs – 6.45%

It will not go away – 64%


Majority do not think IT is doing a good job and is a laggard when it comes to technology, however most do not think IT department will get eliminated either.  IT may be on the path of becoming like HR department – a necessary evil that is universally hated. 

What do you think – add your comments.


Why Verizon got it wrong…

The content is modified to fit your screen and for language for your comprehension… – The Agency :).

Short lazy post on what Verizon got wrong in its Cloud strategy. As always read it at your own risk. (Moreover this is based on zero actual customer interviews) – this is pontification that tries to convince you as if its reification.

Read this first: http://www.asymco.com/2010/10/28/re-framing-the-dichotomies-open-vs-closed-vs-integrated-vs-fragmented/

Italics are from the referenced article.

Interdependent architectures are the only way to rapidly and competitively improve the performance of not-good-enough systems. Modular architectures are the only way to squeeze profitability from more-than-good-enough commoditized technology.

Hasn’t compute and storage become more-than-good-enough? I would argue it has. Look at recent price drops.

Interdependency is often required to raise the performance of a new solution. In contrast, modularity is required to lower the cost of an over serving solution.

Compute and storage are not new services. Verizon might be targeting outside of compute and storage, but not sure if they know what they are targeting yet – I couldn’t tell.

Modular architectures are often the solution when interdependent architectures prove too costly.

Bingo. We are reaching a point where if you have large workloads, AWS is too costly and it is better to go with your own Cloud infrastructure or negotiate special enterprise pricing. The way you beat that is with modular architectures. For ex: using OpenStack.

So the winning strategy depends on detecting where a product lies in its march up the performance trajectory. Before it’s good enough, interdependent systems win. After it’s good enough, modular architecture wins.

I argue the compute and storage clouds are good enough. AWS has won in last 8 years, but it has become beyond good enough – what wins in next wave is modular architectures.

This is where I would put my dough on OpenStack or other open IAAS.

How to disrupt AWS

Post formatted to fit your screen and content modified on the fly for your convenience – The Agency.

What are some reasons why AWS needs to be disrupted?

  • Closed system – everything inside it is magic and secret – what if they are sharing everything with the US govt?
  • Monopolies are not good – no other vendor has their shit together as a result, its AWS and no one else in the leadership quadrant – its not AWS fault that all others are impotent companies, but we as an industry would be better off if there is worthy competition to AWS
  • Pricing – AWS may be price gauging at the higher end of usage – yes, you can negotiate enterprise pricing etc, but many of the enterprise who start using AWS and increase its use across departments ( albeit fragmented) may be getting ripped off. AWS can do it because they are a monopoly and we don’t have an option.

How to Disrupt AWS

1. Established companies can not and will not be able to do it using current structure
2. Needs to form a dedicated BU with its own limited sales team or spin out a new company
3. Do not build from scratch – leverage something with momentum like OpenStack ( or others )
4. Pricing – offer compute and storage for free – make it free for development teams. This is where it helps to have deeper pockets in terms of funding.
5. Offer subscription service for support – nothing new here, but don’t optimize for revenue here – it is just a safety feeling for enterprises.
6. Do not match AWS features before you totally dominate the developer market.
7. Do not attack AWS – do not position yourself as being in a David vs. Goliath fight – position it as your developer outreach play, not Cloud play
8. Focus on developers, developers, developers
9. Be engineering driven- dominate OpenStack ( or whatever IAAS you pick as your base) submissions – in fact your company should consist of 75-80% developers
10. Be Cooler – be the one company that everyone likes when it comes to using Cloud in dev teams.

You will not make money in first 2 years, but once you dominate developer market, you can start adding value added features. You can’t put cart before the horse though, stay patient and stay disciplined.

Unrelated – take a minute to fill out IT Trends survey here: http://www.surveymonkey.com/s/Q8TYTC8

IT Trends – Read at your own risk

This blogpost has been edited and formatted to fit your screen and language changed mid-stream to improve readability – The Agency, funded by your tax dollars.

Stream of thoughts on whats happening in IT – no firm recommendations – read at your own risk, there are no silver bullets here and it may not even enhance your knowledge much.

1. Consumer Software is now more advanced than enterprise software

  • UX – UX of consumer apps like Starbucks, Human are far more better than SAP, Oracle or PeopleSoft. ( I know Apples and Oranges – but, little known fact – when I go to store, I do decide between Apple or Orange for fruit ).
  • Context – when I use my Starbucks app, it knows my context and opens up my e-card to pay.
  • Integration – Starbucks app is better integrated with iOS Passbook than any of the enterprise systems integrating with AD
  • Functionality – Most consumer apps have exact functionality users use – not much gold plating or feature bloat
  • Smarts – Consumer apps are far smarter than any typical enterprise app.

The impact of consumer apps is on expectations – employees need systems that be signed onto, start using in matter of minutes without having to wait for some IT lord to approve it. Secondly, they want it accessible from everywhere, without having to suffer through VPN junk.

2. Enterprise Companies still stuck in old models, not in tune with today’s new employees

As companies start hiring more younger employees, they demand an experience they are already used to. BUs will demand it from IT and if IT can’t deliver, they will try to do it themselves patching together bunch of consumer tools ( think Dropbox + gmail + one of many free conferencing tools + twitter )

Its a reality that most CIOs of the day grew up in the days of green screens and for them web is a huge improvement. What they view as innovation is what today’s ‘younger’ employees view as bare minimum.

There is a huge mismatch in experience, expense and time to procure something using a consumer app vs. using an enterprise procurement system. For ex: compare process of buying coffee at Starbucks using the Starbucks app vs. procure a box of pens at work.

3. Enterprise IT still believes in control based objectives

Typical enterprise IT views employees as someone to protect the enterprise from. Monitoring to stop bad employee bad behavior. It really should be about how to unlock the productivity. If you have to worry about employee’s, how good is your hiring process? Don’t put barricades to make up for your HR short comings.

4. IT Budgets will shrink and they must shrink

IT has become bulky, just like most govt. are. It has structured itself around outdated processes and falsely believes its job is to follow a process defined a decade ago. Innovation takes backseat, while established vendors keep feeding them BS about sticking with what works.

As budgets start shrinking, you will find your enterprise vendor salesmen will start seeing less of you, you will be forced into evaluating cheaper alternatives – you will feel lonely. Get a jump start on it by evaluating more cost effective alternatives – for ex: instead of buying new hardware, evaluate if you can use elastic infrastructure from AWS.

5. Cloud does not solve any of these problems, but if you want to solve them, you need Cloud. ( You also need whole lot of SaaS )

Cloud does not solve any of the identified problems – the solution to those problems starts with IT radically remaking themselves, and in some case, the only positive outcome is for IT to fire itself and decentralize it across different business units.  To fix some of these problems, IT may have to throw their existing processes out the window. This could mean you no longer follow your carefully designed 10,134 step server provisioning process. Also, to improve responsiveness, you may let BU’s provision their own instances of applications ( software, servers etc). This will lead you to Cloud. Now,  you may find that before you get to cloud, your BU is already using AWS. Do not resist it, instead figure out how to adopt it and extend it across the company. Try to add value.

Btw, AWS is not the solution to your problems – it is just one tool, it makes it easy to get started – but it starts to fall apart at scale due to increased costs. AWS is not optimized for massive usage, unless you negotiate special pricing. If you are paying off the street prices and using it heavily, you are getting screwed. So, if AWS is not the final solution, what should you do? Look at an alternative IAAS solution like OpenStack that gives you flexibility on location of your resources and you have better control over costs. Moreover, despite all the API issues, a nice ecosystem is building around OpenStack. Its here to stay.

Should you use PAAS like OpenShift, CloudStack or CloudFoundry – no, not yet – the jury is still out. Unless you are a bleeding edge, stay away from PAAS for now – give it 12-18 months to mature.

Update: Need to read up on CloudStack :).

Unrelated – take a minute to fill out IT Trends survey here: http://www.surveymonkey.com/s/Q8TYTC8