Cloud Change – the Winter freeze

Ok so I am no ITIL practitioner, however I have over my small amount of years in the industry been exposed to a large volume of both internal and externally driven IT Change freezes. Change freezes in the Cloud Computing world I think are going to transform in a very big way. When thinking about Change and how this is going to work in the new generation Cloud I’m lost on what dramatic force Cloud delivery models will have on the traditional change freezes that occur at the festive period…I do have some views to share so I have compiled this post to articulate them.

Nostrildamus begins his prediction

So lets start, my first view is that if change freezes are not dissolving already today within Infrastructure landscapes that embrace the new wave of Cloud computing methodologies then they most certainly will do over the next 5-10 Years. To explain why I think this, I will break down a few simple reasons. My views focus on changes that occur within the Infrastructure layer or more aptly IaaS, I do not discuss other layers of a cloud computing stack such as PaaS and SaaS layers.

Public Cloud

Infrastructure environments within this area include a consumer base that generate the  most profit and revenue at the typical festive freeze period and in turn actually demand more Infrastructure capacity and dynamic growth. It is the biggest time for actually growing customer base, consumers will buy the most endpoint devices that connect to the services hosted on IaaS i.e. Internet, Smartphones. Also anyone who is an existing customer will use more services during the busy period, so I think you’ve got my drift that “were gona need a helluva a lot more power captain”.

This large demand for consumer cloud compute to support the larger influx of end user population usage, designing for this peak demand in public cloud service is not something that can be forecasted as well as some would like, and without doubt it is the area of Cloud computing that makes use of the emergent “Cloud bursting” under high utilisation peaks.

My view is that “Cloud burst” cannot be achieved when stringent change freezes are imposed on the IT landscape that operates it, cloud burst does not introduce change per se, however it merely increases the logical resources that run on top of physical devices which to someone whom is within a non Cloud mindset would assume this is a change and in turn may effect running production.

Architecture within a private cloud needs to be highly available and able to sustain any potential outages upon underlying architecture, this may be provided by either Infrastructure level technology or by the running application being able to scale accordingly across stateless nodes. This highly available service driven environment enables a potential of being able to reduce the need to freeze any changes, however this process needs to be completely invisable to end users and responsible custodians of the relevant environments.

Verdict - Change Freezes that may be imposed on landscapes supporting Consumer Cloud will be the first that gradually dissolve, in some businesses they may have no choice!
On premise virtualised and automated shared pool of resource – Also known as “Private Cloud”

You may notice that I have not named this delivery model as “Private Cloud”, today most organisations mainly just have a private on premise virtualised and automated shared pool of resource, and are not yet at the holy grail stage of the “Private Cloud”. Unless an organisation who is working towards adopting a private cloud was extremely efficient before adopting the cloud strategy, it is more than likely that a strategy of Private cloud is not anywhere close to achieving the same level of decouplement, fluidity and efficiency as a typical IaaS based public cloud service do when introducing change and performing service fulfillment.

Changes at festive periods within a Private organisation also exist due to lack of staff resourcing, will the touted benefits of hands off automation and workflow for service fulfillment being allocated from a pool of resource really need to warrant or be subjected to change freezes? You could argue that as with public based cloud services cloud burst or service fulfillment through a web portal for end user requests is not a change so that relevant project or business activity would not need to schedule around a change freeze.

Unfortunately though, with the lack of full maturity of Private cloud in most organisations I expect that this means change freezes are going to certainly have more shelf life than consumer or public clouds with the main reasons being due to the way that service management  teams internally are still acutely aware of the Infrastructure landscape that operates the service currently touted as Cloud.

In most shops the complete underlying Infrastructure has not yet transformed into the new architectural design model that is orchestrated and workflow driven. Change freezes and other related process won’t be able to dissolve until the services and potential service disruption is completely invisable to the end user whom is either a user or responsible for the underlying Infrastructure landscape.

Verdict - Change Freezes within internally operated environments will take far longer to fade than public based clouds due to the legacy it will need to transform away from.
An off premise hosted virtualised and automated shared pool of resource - Also known as “Private off premise, Hybrid or Private cloud”

Ok this form of cloud delivery is introduced in the same sarcastic way as i’ve used with describing the on premise cloud and this is mainly due to again a lack of maturity. As with Private on premise cloud, this area of IT service delivery is currently at an early stage of adoption by most Service Providers and is at the early stage of being moved into core portfolio offerings for customers. Service Providers see Cloud models as the next big opportunity to streamline and make up for the lost ground and revenue made within the Virtualisation era so the future vision to make outsourced hosted environments “Cloud” based will most certainly become the de facto.

As with Private on premise based Infrastructure pools, potential Change freeze fade within the Service Provider world are also very much based on first the level of maturity that the SP has for service fulfillment and automation. and secondly and more importantly, having the capability to operate effective multi-tenancy and a complete decouplement of a shared resource pools between customers, and don’t forget this is not limited to Server workloads, this is relevant across the complete IT hardware stack end to end.

Again resourcing is a reason for change freezes within hosted environments, this also waterfalls down to the plethora of sub contractors and contractors that the incumbent supplier would use for customers. If the future of cloud is to be a self service based affair from shared resources why would the engagement and fulfillment process that is experienced today within outsourcer’s need to exist and be limited over festive change periods in years to come? I suspect a large legacy of people within organisations will still remain, and at the end of the day customers like to know they are getting value for money by a physical bod fulfilling either design or build.

Verdict - Until Service Providers completely make underlying environments that support customer workloads and services completely invisible, Change freezes will still exist within this model. Additionally Outsource providers have a much larger broad range of diverse customer requirements which cannot be transformed globally overnight, this will reduce the time that examples processes such as Change freezes take to dissolve.
Post Summary

Hopefully I have articulated well enough to you to justify why I think Change freeze days are numbered within a Cloud based world.  Timeframes of this likely occurring is way over 5-10 Years and this will mainly be due to a legacy and generation of practices that have been implemented into organisations and also a an existing Infrastructure that is not yet capable of delivering new additions or new changes in a non disruptive manor.

Also lastly, in order to succeed in providing non disruptive service to your business you will need to either first write off financially your existing assets/agreements or find a way of actually funding an Infrastructure landscape capable of supporting a Cloud, something that unfortunately most companies will be limited too for many years to come.

Mcafee – Enters the Ring!

After speculation I made last month about why Intel has bought Mcafee, I think the announcements on the new ‘Sandy Bridge’ architecture may now starts to make the picture clearer as to why.

With a more consolidated on chip strategy being developed by the little white men in dust free suits, are we potentially going to to see a nice big fat Security box within the ‘Sandy Bridge’ Ring alongside others? Who Knows…

vCloud Director – the vmworld koolaid hangover

Ok so you have been to VMworld 2010, got out of bed at 730AM for the keynote on vCD with before hand having a ridiculous burrito and stale awful coffee or such like for breakfast at the VMworld catering, before rushing to get into the first sessions feeling like you are merely a single sardine in a sea of many through the crowds and then after all that got home after those general product sessions to install and see vCD for yourself to find;

  • It requires Redhat 64 Bit,
  • Needs an Oracle DB
  • Has no OVF based appliance

Disappointed comes to mind with a prereq list like that, especially after i’ve had a twitter stream, RSS feed and work email box inundated with hype over this product. I would expect with the push of virtual appliances and the alliances with Novell this could be packaged into a VA so at least simpletons like me could at least get half way to deploying.

It’s a shame, it appears vCD that I know full well has been in development for sometime under the Redwood guise has merely been released to woo customer and raise share prize at VMworld.

Well I will tell you this…if a product that is going to “take me into the next decade of IT” as the marketing waffle suggest it has got a lot of work to do in a short space of time to convince me of this…..personally releasing a product that is as fundamental to VMware’s future as where the initial ESX releases back in the original rise of VMware I think they are treading on thin ice.

End of rant