Blessay: Public Clouds Wont Be Big in 2014, Private Clouds Will – A Contrarian View

It’s becoming clear that 2014 is going to see a significant push and adoption on Private Cloud supported by marketing and technology from incumbent vendors who are starting to lose substantial revenue. Public cloud companies don’t spend with Cisco, IBM, HP or other recognised manufacturers and it’s time for them to defend their turf. Will this change

Incumbent Vendors lining up Against Public Clouds - There are number of early signs that the incumbent vendors are realising that their product revenues are under serious threat by public clouds. Cisco, VMware and Redhat don’t make money when public clouds take over applications and their long-term revenue is under threat.

So the vendors are beginning huge marketing campaigns to encourage corporate IT to move into private clouds. This will have the effect of preventing public cloud adopting.

Enterprise IT is Big - and much bigger than public cloud of today. There are breathless predictions that Amazon is expected turn over $9B in cloud computing next year and that business took just “ten years” to build. But Cisco turns over $20B per annum and took 20 years to build so Amazon looks more pedestrian by comparison.

Extend the comparison – VMware turns over $5B per annum, Microsoft does $80B, HP is $113B, Brocade $2.2B, Juniper $4.5 and IBM is delivering $101B a year.

I suggest that if these companies commence collective marketing efforts into promoting private cloud then much of the momentum into the public cloud will disappear almost immediately.

Clouderati & Echo Chambers - We are seeing regular commentary that high-profile figures involved in promoting the public cloud are living in an echo chamber that leads them to conclude that the only cloud is public. I have a contrarian view on public cloud that it might end up like HDTV – nice, desirable but there is not enough content (applications) for widespread adopting and too expensive for most people to use.

There as vast numbers of people and companies that are not able to speak publicly about their IT functions but infrastructure vendors will tell you about market research that shows real cloud growth.

racks-of-computers-cloud-ready-595Public clouds are proving grounds for Private Clouds – One interesting side effect of public clouds is that they are the really effective proving grounds for technologies that private clouds can use. Overlay Networking, Puppet/Chef system orchestration, NoSQL databases, IP Storage, Virtual Appliances  are just some of the technologies that have jumped over the usual process of “give it a few years to wait & see” and moving rapidly into adoption because there are large reference sites in Amazon or Rackspace.

My take is that public clouds have been valuable test beds for low quality companies to adopt innovative technology and achieve a business advantage. After five years we are able to see which technologies are suitable for use in the corporate data centres and enterprise deployments.

Power problems  – the public cloud companies are facing supply problems for power in many countries. There are a lot of articles about alternative power generation that strongly signal power shortages for new data centres. Note that distributed nature of alternative energy does not bode well for big data centres.

Low Power Infrastructure – this is coming from several sources. There are servers, storage and networking equipment that can operate at room temperature (up to 60 degrees) and thus don’t need cooling. Intel is working on new technology that consumes much less power while increasing compute output. ARM CPUs have potential to shed certain types of integer-based compute into low power.

I currently have the view that data centre infrastructure will get physically smaller and consume less power at a rate that exceeds the capacity increases. The adoption of cloud technology will increase application over the next few years.

This will drive existing data centres to have extended lives and delay any forced transition to the public cloud. The backdrop of security and legal issues in the public cloud have been highlighted by NSA/Snowden and large breaches like the Adobe breach where 135 million customer credentials were stolen ( but ONLY 38 million were current users..) make it ever harder to move services externally.

Flow On Effects –  There are number of other flow on effects from public cloud. Today, I think that the most interesting is the radical price cuts in infrastructure hardware. Servers, network and storage are being commoditised at an unprecedented rate and headline prices are falling quickly.

Just a year ago, building a 10 Gigabit Ethernet network for an Enterprise data centre cost about $5000 per port. Today it is $500 per port and still falling. Equally, storage arrays are being assembled from x86 based servers. And customers are beginning to turn away from blade servers to rack mount servers for better virtualization density.

And finally, cloud management platforms like OpenStack are getting mainstream support and coverage. The software ecosystem around cloud management is receiving vast amounts of funding and growth.

The View – I like public clouds as a business model and respect the value that can be delivered their focus to using technology in creative ways. As a technologist its tempting to get excited about their ability to adopt and deliver new technology.  And there are definitely used cases for public cloud to host services for everyone. But I wonder if that technological fascination is like goggling at a car accident when  you are thinking “what would I do in the that situation?”.

Today, Amazon is a poster child for new business models like Dropbox, Netflow and Instagram. Yet none of these companies  would exist without large quantity of pre-built infrastructure that could be rented monthly and increased if they got lucky with their customers. Private clouds are much more useful. They can be customised while costs are fixed, predictable and manageable. The long-term value of public cloud is significant but I don’t see any proof that private clouds or static IT will be replaced in the next 20 years.

Who remembers outsourcing fashion from ten years ago ? And who has noticed that outsourcing is largely over and being considered as just another option. Apply the same logic to public / private clouds, we will have both and customers can choose the best solution.

And with a little pushing and encouragement from the incumbent vendors, who have been trusted advisors for decades, I think that most companies will rapidly ramp their internal adoption.  I think Private Clouds is the next wave, what do you think ?

About Greg Ferro

Greg Ferro is a Network Engineer/Architect, mostly focussed on Data Centre, Security Infrastructure, and recently Virtualization. He has over 20 years in IT, in wide range of employers working as a freelance consultant including Finance, Service Providers and Online Companies. He is CCIE#6920 and has a few ideas about the world, but not enough to really count.

He is a host on the Packet Pushers Podcast, blogger at EtherealMind.com and on Twitter @etherealmind and Google Plus

You can contact Greg via the site contact page.

  • Marcus Gentile

    I work for a highly-security conscious company. There are very few services we can or are willing to entertain delegating to a public cloud due to data loss/exposure concerns. Certain publicly available info, such as the company website, are easy to shift over to 3rd-party IaaS. Internal R&D, customer data, and security services are another story. Private cloud, or just internal hosting, is the only way to go for that sort of stuff for now. It seems like public cloud will become just another tool in the IT toolbox rather than “the Holy Grail” of enterprise IT.

  • Denis Borchev

    Could you please tell more about “And customers are beginning to turn away from blade servers to rack mount servers for better virtualization density.” (Flow On Effects point) ?
    I find it a little confusing, as part of the blade server’s promise is density increase, I thought.

    • http://etherealmind.com Etherealmind

      Sure. Blade servers are relatively expensive because of all the customisation that is done. The idea of sharing a backplane is allow multiple servers to share resources like FibreChannel and Ethernet adapters. This made sense a few years ago when those technologies were expensive and sharing them reduced the overall cost.

      Today, you can buy servers with FC & Ethernet for less than the cost of the NICs back in the day. IN an enterprise environment, there is a use case for bare metal installs of the OS where blades make sense. However, when virtualization is used, blades add no value in terms of features because the hypervisor replaces those functions.

      Invest in the KVM, ESX software and use a cloud manager like OpenStack or vCloud to add new functions. It’s cheaper and more flexible in the short term. Cheaper hardware can always be replaced if something new comes along in the future.

      Blade density is an old idea. Hypervisors make them obsolete.

Subscribe For Weekly Updates by Email

Get a Weekly Summary of Latest Articles and Posts to your Email Inbox Every Sunday

Thanks for signing up. Look for the email from MailChimp & make sure you confirm your email address. You may need to check your spam or gmail settings to be sure of receiving the email.

Note: You can unsubscribe at any time using the link at the bottom of every email.