Why the Public Cloud is a Big Fat Enterprise #Fail

With the advent of the consumer driven enterprise, there are now two ways to define the success of a platform. One is to look at adoption rates and declare a platform successful based on how many people are using it. By that measure, the public cloud is a smashing enterprise success.

Billions of users worldwide are using public cloud services such as Dropbox, Google Drive (formerly Docs), Skype, Evernote, YouTube and, yes, even Facebook to do business with colleagues, customers, partners and vendors. 

But the other way to define platform success is to look at what businesses, organizations and institutions are actually investing in and encouraging people to use. By that measure, the public cloud is a resounding enterprise flop. Every one of the public cloud services I just mentioned top the list of blacklisted apps in the enterprise. And that list keeps growing.

Moreover, adoption of Google Apps appears to be sputtering, despite reports from the company. Still with the beefed up version of Google Docs, Google Drive, now available, Google Apps is certainly one of the highest profile public cloud platforms.

Now, you could argue this shows CIOs and their IT teams are clueless, and they prove that by saying “no” and pushing back against consumerization in the enterprise. Rather than use Dropbox, IT insists on using Sharepoint. Rather than use Skype, IT requires WebEx. And so on.

I would argue, however, that it’s the public cloud services vendors who are, in fact, clueless. Large enterprises continue to embrace private clouds. IT organizations increasingly understand the risks, opportunities, roles and potential benefits of public and private cloud computing. And they’re largely putting their chips on the private cloud card.

If I were running a public cloud service, I’d be worried about IT blacklists and private clouds used to meet requirements that could be served more quickly and cost-effectively by a public cloud solution. I’d be asking IT leaders why they’re attracted to private clouds and repelled, overall, from the public cloud. And I’d start transforming my offering to give IT what it needs.

Enterprise IT leaders want tight security in public cloud services and SaaS apps. They want to ensure compliance with HIPAA, FISMA, SOXand other regulations. They need enterprise data encrypted at rest and in motion. They need directory services integration. They need disaster recovery. They can’t afford high latency, downtime or a single data breach. The list goes on.

Let’s use Amazon’s AWS public cloud as an example. When you shut down a workload, that workload is gone and is unrecoverable. If a user accidentally shuts down a workload, everything they were working on vanishes.

From Amazon’s perspective, that’s your problem. You’re expected to have backup and disaster recovery systems in place. You’re expected to design applications for failure rather than design for continuity. You have no control over continuity in the infrastructure, so you have to code it into each and every public cloud application.

Simultaneously meeting the needs of consumers, while addressing the needs of IT leaders, is exactly how a small company named Microsoft evolved the original consumerization of IT — personal computers running DOS and Windows — from a rogue device into a primary platform.

Public cloud technology vendors would do well to reuse that model. Giving consumers what they want, even to the extent of providing tools and support to help users bypass their IT departments, might be an effective way to quickly grow a user base. But it’s a dead end over the long haul.

Certainly consumers’ needs should be met. But give the IT organizations that support them the tools they need to provision, integrate, authenticate, secure, manage, automate, monitor and audit. Instead, however, the majority of public cloud vendors are counting on the consumer driven enterprise to force-feed their solutions into the enterprise.

The growth of the public cloud proves consumers want change. The growth of private clouds proves IT is open to change. If we don’t want to see the blacklists get longer, public cloud vendors must start listening to IT if they hope to succeed in the enterprise. By coming together, the cloud overall becomes more productive, safe, and effective for everyone.


The following two tabs change content below.

Andi Mann

Andi Mann is vice president of Strategic Solutions at CA Technologies. With over 20 years’ experience across four continents, Andi has deep expertise of enterprise software on cloud, mainframe, midrange, server and desktop systems. Andi has worked within IT departments for governments and corporations, from small businesses to global multi-nationals; with several large enterprise software vendors; and as a leading industry analyst advising enterprises, governments, and IT vendors – from startups to the worlds’ largest companies. He has been widely published including in the New York Times, USA Today, CIO, ComputerWorld, InformationWeek, TechTarget, and more. He has presented around the world on virtualization, cloud, automation, and IT management, at events such as Gartner ITxpo, VMworld, CA World, Interop, Cloud Computing Expo, SAPPHIRE, Citrix Synergy, Cloud Slam, and others. Andi is a co-author of the popular handbook, ‘Visible Ops – Private Cloud’; he blogs at ‘Andi Mann – Übergeek’ (http://pleasediscuss.com/andimann), and tweets as @AndiMann.

This article has 17 comments

  1. Very interesting article, altough I found the title a little bit misleading and (maybe purposely) provocative.
    The way I see it public cloud will dominate the enterprise space in the medium term future the same way that off-premises
    datacenter services have become common-place nowadays.
    Security and regulations are a barrier for sure but once CIOs and CTOs fully understand the gains and advantages of using
    a public cloud and realise that cloud providers are way more secure than most on-premises installations, this will be a no turning back path.
    As for an user accidentaly shuting down an unrecoverable workload, this can and will happen no matter if one is using a public cloud or not.

  2. Thanks for the comment Rodrigo.

    Yes, the headline is (intentionally) provocative, but I really do mean it. I talk to so many enterprise business and IT leaders that really want to be *like* public cloud providers, but absolutely don’t want to *use* public cloud providers. This is a massive disconnect, and CSPs are fundamentally failing to deliver what enterprises need.

    I would take issue with a couple of your points, but agree with some too.

    I dispute the general statement that ‘cloud providers are way more secure than most on-premises installations’. Plenty of cloud providers have had horrible breaches; plenty of on-premise data centers have never leaked a byte. Such an evaluation is definitely case-by-case.

    And my point about losing data just because of a server shutdown is that this absolutely *won’t* happen anyway – it is quite reasonable to assume that enterprise IT includes backup as part of a standard server build; the same assumption is generally not true of most public cloud providers.

    Still, I definitely agree with you in some degree. Your outsourcing comparison is particularly interesting, because I think it is very appropriate. Indeed, we are repeating some of the same mistakes (especially the blind trust they will do ‘the right thing’ without any contractual requirement to do so). Yet like with outsourcing, I agree that acceptance of public cloud will rise over time, as you say.

    But I really do believe that CSPs have to fundamentally change their attitude to enterprise requirements – not just security, but also governance, continuity, policy, SLAs, financial mgmt, performance, and more – before that change will happen.

    Thanks again for taking time to comment!

  3. what a great blog to promote software.

  4. Nice post Andi. Provocative? Yes. Bold? Yes. Fair and well reasoned perspective? Yes again.

    Now, do I agree? Hmm… yes and no. Your post covers a ton of ground, including consumer oriented cloud services. Since I have more direct knowledge on public infrastructure-as-a-service (IaaS), I’ll focus my comments there.

    I’m seeing a mix of success and fail for enterprise IT. Starting with success, I’m not seeing deep success across the complete enterprise IT portfolio. What I do see is a bunch of mini success stories with test and dev environments, faster starts to some software development projects, some interesting green-field cloud applications, a few migrations production applications that fit well, and so on. These are real successes, sometimes with quantifiable value and/or cost savings. Further, enterprise IT seems to be gaining more and more traction – sometimes by using emerging tools for application migration, hybrid cloud connectivity, etc.

    What I don’t see from enterprise IT is reactions like, “OMG, public cloud is so much better that we need to move 25% of what we have there or we’ll lose out to competitors that do.” They love the automation and – as you mentioned – want to have that in their own private clouds (or even as part of their more traditional environments). [Note that I’m ignoring the 1% club including the likes of Netflix and other Web / Internet businesses with single, massive, scale-out applications etc.!]

    To overstate it a bit, I think it largely comes down to enterprise IT having loads of applications that aren’t easily moveable (and then manageable) to public IaaS environments and that don’t necessarily fit well in public clouds. They sure weren’t designed to run in a public cloud. Then, when you look at the specific issues they face, you can see how pragmatic enterprise CIOs really are. Yes, a few are simply cloud haters… but the majority seems to be quite in touch with real public cloud hurdles. This is where I imagine we may both see something similar.

    > But give the IT organizations that support them the tools they need to provision,
    > integrate, authenticate, secure, manage, automate, monitor and audit.

    That is a great list of some of the gaps enterprise IT faces. Interestingly, I’m seeing *huge* progress on all of the above. There are great products and services available (sometimes directly from public cloud IaaS providers, but often from third parties – including CA, I might add) that do all of those things. OK, even when you pull all those great add-ons together, there may still be gaps… e.g. the public cloud simply won’t support certain network configurations, certain hardware architectures, etc. [though I’m just starting to see some good progress on these items too!]

    But don’t get me wrong… While I see what I believe is a clear trend to address the gaps you’ve listed, I’m far from saying they are resolved. Or even from saying it is “worth it” to work through them in many cases. For many applications – particularly the vast majority that weren’t designed for most of today’s public clouds – it can get awfully expensive. And for applications that are working perfectly well otherwise, why move them to a public cloud in the first place? There needs to be clear business value.

    I guess I should net it out… ;-) In terms of overall public cloud readiness for enterprise apps, we aren’t there yet. While I can’t get myself to say “big fat fail,” I think that so far it is more fail than success for the enterprise. Yet, success is growing… 1) more enterprise apps are becoming cloud ready and 2) more public clouds and 3rd party tools are making it easier to support existing enterprise apps. Regardless of the exact semantics, it makes tons of sense to call out these issues. The enterprise IT leaders and CIOs who are not aware of them will surely face a big fat fail!

  5. Thanks Paul, really great comments, and very informative. Just found myself sitting, reading, nodding along. Greatly appreciate it. :)

  6. We’re certainly having different conversations with the senior IT leadership in some of Australia’s largest enterprises. They are amazed, surprised & blown away by the cool things that can be done with AWS. They are experiencing dramatic improvements in agility, flexibility, scalability and improvements in control and security.

    Well architected environments, built with a clear understanding of what’s possible with AWS are making a huge difference to execution capabilities of enterprise IT all without the need to spend large amounts of capex on interim private cloud solutions!

    Of course the right workload for the right platform, all hosted, private, hybrid, public, IaaS, PaaS, SaaS discussions are shades of grey, no one person or argument is right.

  7. It’s an interesting viewpoint. I’m not entirely sure enough time has passed to make that call globally across all enterprises. We enterprises tend not to move very fast with broad adoption of anything so it’s hard to gauge yet whether the public cloud adoption is a miss with enterprises. Throwing out a few of the very large enterprises I think security, data safety, backups etc have been more thought through with some of larger public cloud providers versus your average enterprise(Notice I said average, not elite few). I know from my enterprise viewpoint we are expanding our public cloud usage but I wouldn’t call it a sprint pace.

    The point about compliance is probably the single biggest gut check most CIOs as well as architects will tend to have when pondering public clouds. It’s really not entirely up to the cloud provider to provide that, it’s more of a joint effort, but it will still give CIOs pause.

  8. Rakesh Malhotra
    Friday 1 June 2012, 1:32 am

    I think public cloud will be part of the overall IT portfolio and the end state is hybrid (private/public/IaaS/SaaS). With that said, we definitely hear from all of our customers that regardless of how things turn out, they need to run their own datacenters like a world class service provider. There is definitely a certain degree of presumptuousness and naïveté in the cloud echo chambers about how enterprises operate and how to satisfy their requirements.

  9. Hi Andi,

    Agree with you wholeheartedly on the IaaS side, but wonder where you stand on the enterprise SaaS side.

    Salesforce has already made significant inroads into the large enterprise, and we’re starting to see Workday and ServiceNow displace traditional on-premises vendors. Netsuite not far behind.

    I would consider all these public cloud providers. Clearly not in the same category as the consumer apps you listed above, but also clearly not private cloud. I would consider these ‘enterprise’ public cloud providers – sales pipeline focused on corporate customers, but architecture is based on a multi-tenant platform with standard terms, SLAs, etc.

    I think we’ll continue to see a disconnect here on the apps front – IT organizations building private clouds, line of business leaders buying public cloud SaaS solutions, without much forethought as to how these will work together.


  10. As always, great thought provoking stuff Andy. What’s interesting about cloud is that many are building them and using them without a clear definition of WHY. It is just the cycle-time for using internal resources? It should not be cost for large enterprises as they have the opportunity to build private clouds that have similar economies of scale as AWS and build in their own assurances tailored to business or application needs.

    There is not a single answer. I just feel that it shouldn’t be difficult to answer the question. Using public cloud? Great, why? What is the driving reason and where is the analysis backing up the decision?

    There is probably a healthy dose of lazy involved. “Im using public cloud because it was easy. I have not bothered to examine the fit for my application or take the time to understand any potential compromises in manageability or security.”

    Wait until there is a fire and then solve that problem. Rise. Repeat.

    To be clear, I’m not pro/con any cloud model and believe there is a fit for many enterprises and applications for all variations. But I’m against blind group think that “Cloud == Good. Nuf’said.”

  11. Mark, thanks for commenting. I especially liked this:

    > “enterprises tend not to move very fast with broad adoption of anything”

    That is certainly true, and we witnessed VM Stall (the slower-than-predicted uptake of virtualization) for the same reason. Maybe ‘Public Cloud Stall’ is just another symptom of the deliberate nature of enterprise technology adoption, rather then inherent failure of the cloud service providers?

  12. Andy, thanks for commenting. I certainly am not saying that large enterprises are no adopting public cloud at all. I absolutely do believe in the benefits of public cloud. And there are definitely many exceptions to my observations – in use cases, providers, and enterprises.

    But the data is clear, private is outstripping public, and the majority of enterprise CIOs and IT execs remain leery of public cloud (saw new data on this today, actually).

    The exceptions you mention are real, I have no doubt, but are anecdotes that still do not disprove the data. Maybe we are talking to different enterprises – although I know Australia is a different kind of market to the US.

  13. Rakeshm thanks for your comment. I cannot agree more – for almost all large enterprises, the hybrid cloud is certainly the end game, and a very smart choice at that. We even see that in some new-world enterprises like Netflix and Zynga, who both run hybrid IT environments today.

    And the more we can do to listen in on what is happening outside the “cloud echo chambers” you so accurately describe the better! That is why I try to spend as much time as possible talking with actual IT practitioners – leaders and others. It is otherwise way too easy for me to believe the hype!

  14. Hi Stanton, great to hear from you, and thanks for the comment. SaaS is an interesting discussion itself – and yes, I was definitely focusing more on IaaS and PaaS in my original post. SaaS is really an exception to this position, though possible not completely.

    Salesforce, Netsuite, ServiceNow, Intuit are all doing great enterprise business – and I would add CA Technologies to that list, btw, as we have a wide range of management solutions now available ‘as a Service’ (Nimsoft, Service Desk, ARCServe, Clarity PPM, AppLogic, CloudMinder) which are doing great (contributing to us being #2 in IDC’s ranking of cloud system management providers).

    So SaaS could well prove the exception to the rule. Though I do think many (not all) of these are still failing enterprise IT demands – in areas like assured performance, integration with back-end systems, governance and compliance, etc.

    But as you say, this is increasingly not holding business buyers back – even if IT is not fully satisfied, the business sees these solutions as ‘good enough’, while also being faster, easier, and cheaper than their own IT department.

    Which sets up a whole different discussion – on the pros and cons of rogue IT – but I will leave that to another day :)

  15. Hey Mike, great to talk with you again – it has been a while! Thanks for the comment.

    You make a great point, and I think it ties in a little with Mark’s too – the deliberate choices are one side of the equation, but the poorly-considered choice (no matter how otherwise apt) is the other side. I see a lot of orgs assuming cloud will be the great solution they need – to issues variously of agility, cost, security, continuity, scale, and more – but then realize that it is really just an enabler. Garbage in, garbage out.

    I am a big fan of planned adoption of cloud, especially the need to look at what cloud could do to transform your existing applications. This may be a shift to IaaS, PaaS, or SaaS, or it may be keeping some apps as ‘cloud-free’.

    Without a real analysis, how can any org possibly make intentional decisions about cloud adoption? And without intent, you are indeed setting yourself up for fire drills, if not outright failure to achieve objective (assuming they have even been articulated).

    Like you, I am not pro- or anti- either version of cloud. And CA supports as many varieties of cloud as our customers need! But I do see room to improve in both areas.

  16. Good job, Andi.

    The consumer / enterprise schism puzzles me. Did some legendary Silicon Valley VC long ago drunk-rant at a cocktail party “Consumers *hiccup* or enterprise. Techsss pick one or lose at boooooth” ? ;-)

    Even a service such as YouTube could have a plausible pay per seat enterprise business along with its ad-supported cute animals, stunts gone wrong and music video consumer site… as long as it provided a few software tools to help video-ize the modern knowledge worker and work. Save money or make money or … black list for you!

    The precedent isn’t such a reach. Phones in the home. Phones in the work place. Same basic service, two-way talk path, as a starting point. Different grades of service and feature sets and pricing models. Phone company would even sell you a switch if you wanted to be your own mini phone company. Worked for how many decades?

    IMO, these web-based services providers (cloud, wind, fire, dust, whatever) would be wise to free themselves from the myth that the world began in 1993 and nothing can be learned from the tech that stood the test of time long, long before then. Cloud vendors too. Cloud vendors have a once in a generation opportunity not to define “cloud computing,” but to redefine computing itself (every where for every one on every thing).


    p.s. No consumer cares how many acronyms or buzzwords the cloud-erati invent. No enterprise decision maker will adopt a service because an Eminem music video got 400MM views and “lots of people are doing it.”

  17. Thanks, Andi.

    Refreshing to see a balanced look at rapidly growing “new trend”.

    Please keep posting clear, practical thoughts.

    -Jeff in Northern Los Angeles County, California, USA

Leave a Reply