Overcast with Sunny Spells and Storm Clouds to Follow
Following yesterdays blog seeking to make a few pointed predictions, I thought I would add some thoughts on what impact this will all have for the future with a short, medium and long term outlook for private cloud and OpenStack:
Short term: people will struggle to make a profit from Openstack
With Liberty (the 12th release) OpenStack is starting to become enterprise ready, but it is still behind AWS on the maturity curve, needs a few more fixes and the skills are lagging the technology – meaning that until the skills catch up and the skills gap closes a little firms will struggle to make a profit from OpenStack.
[easy-tweet tweet=”With Liberty #OpenStack is starting to become enterprise ready” user=”billmew @comparethecloud” usehashtags=”no”]
Admittedly there are a few pioneering client engagements (many of which we heard about at the OpenStack Summit in Tokyo), but most players are still working on a number of PoCs (Proof of Concepts) and are yet to really industrialise their offerings and really reap the rewards.
Medium term: there could be a fortune to be made moving clients to cloud
The gold rush is almost upon us and many firms could strike it rich (at least for a short while). It is all about the applications. These were almost all written to run on Oracle or NoSQL databases. Once a few innovators figure out how to engineer a move from Oracle or NoSQL databases to containers, the flood gates will open and a lot of money will be made in the ensuing migration phase. Amazon AWS already had Oracle clearly in its cross hairs with its recent launch at re:Invent of tools to simplify data and database migration to AWS. Others will improve on this and enable migration directly to containers – which can then run on any cloud platform.
This gold rush will eventually come to an end however – these applications can only be migrated once. It is a bit like VMware that made a fortune helping firms virtualise. It is something that you only do once. At the time VMware could charge a real premium for it and make a fortune, but eventually other cheaper options emerged (like KVM) and the market became saturated. VMware then struggled to add further value (and Michael Dell may find it hard to use VMware to help pay off the debts he incurred buying EMC).
Interestingly mainframe applications may be among the ones last to shift, meaning that, in an ultimate irony, this legacy particular platform may outlast the Unix/RDBMS rivals that tried to replace it.
There is a great deal of IT budget that has been consistently drained in the form of Oracle licence renewals that will then become free for reinvestment in a wave of possible innovation as firms leverage their freedom from the constraints of many of their legacy data stores and seek to capitalise on the potential benefits of containerised environments.
[easy-tweet tweet=”Sometimes the rise of containers is portrayed as competitive with #OpenStack” user=”billmew” usehashtags=”no”]
During this gold rush phase proximity will be important, and firms will want to locate their compute power near their big data stores. This will be a real opportunity for private cloud implementations – a brief period in which the previously mentioned private cloud/OpenStack premium could well be justified. Sometimes the rise of containers is portrayed as competitive with OpenStack, but in the medium term at least OpenStack and containers will co-exist.
Long term: OpenStack risks becoming a backwater
The whole argument for private clouds is that no matter how compelling the public cloud services are, there will always be the following compelling drivers for private cloud:
- Financial advantages (in certain circumstances) – although the private cloud/OpenStack premium will become increasingly hard to sustain
- The need for technical control over one’s compute environment – although increasing usability and flexibility as well as innovations such as containers are making this less relevant.
- Reasons of proximity (for instance, as mentioned above, putting the compute power near big data stores) – although once the gold rush is over and applications are all containerised, proximity will no longer be as much of an issue.
- Reasons of protection, risk aversion, compliance and regulation – although further innovations in security and other area by public cloud vendors and their ecosystem partners are addressing these concerns.
The majority of compute cycles already run on public rather than private clouds (across SaaS, PaaS and IaaS). The share of compute cycles on public cloud will increase meaning that the percentage of compute cycles on private cloud will consequently decrease. Add to this the unavoidable private cloud/OpenStack premium and the increasing relative irrelevance of at least some of the major corporate sponsors behind OpenStack and it is not hard to see private cloud and OpenStack becoming a relative backwater – possibly even before it reaches full maturity.
So where does that leave us all…
This leaves little in the way of a compelling proposition for private cloud at all in the long term. There will be a gold rush and a brief private cloud bonanza – as the industry turns on the likes of Oracle, seeking to break down the barriers of the data store enclosures and free the data – but it will be short-lived.
There will be a gold rush and a brief private cloud bonanza
The challenge for enterprise tech vendors will be in remaining relevant. Will clients continue to pay HP, Cisco, Oracle or indeed anyone else a private cloud premium? Can they continue to provide the same operational and economic flexibility as public cloud, with the control and autonomy that comes with owning your own environment – while reducing the private cloud premium to an acceptable minimum?
The challenge for the predicted ‘handful of players [that] dominate public cloud with their own proprietary stacks’ will be one of continuing to innovate, and to take advantage of economies of scale and fast cycle times in deploying new functionality, and also of staying ahead of what can be achieved just as easily in software.
In the future it won’t be all about public and private clouds any more, but about local and global ones. Local clouds will be as stipulated by national governments (such as the G-cloud in the UK) and global ones will be provided by the usual suspects: AWS, Google, Azure and maybe also IBM (which is keeping pace through its SoftLayer and BlueBox acquisitions), with maybe Alibaba’s Aliyun as well.
The future will be all about software
The future will also be all about software, as everything is becoming software defined.
So lets see how things unfold over time. Please feel free to come back and challenge us on these predictions. We love nothing more than a good debate!