Perot Systems bought last week by Dell and then ACS go this week to Xerox so the question is who will be next?Other big names that may possibly go include;- ATOS
- CSC
- Fujitsu
- Cap Gemini
- Computacenter
- Logica
- Steria
It is amazing how much consolidation is occuring and how quickly in the industry today and its quite alarming what the big powerhouses are prepared to do to jump onto the IT services bandwagon, Especially when you hear about Outsourcing pricing being whittled down to bare minimal profit margin with customers demanding much more out of there contracts.
Now if you were Xerox or Dell why would you buy an IT services integrator and provider? Many thoughts come to mind such as;
- Entry into the IaaS space
- Build a portfolio of offerings which will be ready for the economic upturn
- Improve indirect sale, they will claim to remain agnostic but the longer term goal is to adopt customers to there platforms/solutions
- Quash competition, namely IBM, Oracle and now HP after EDS
Future Predictions
Oracle buy FJS or even possibly CSC, and why do I think this? They both have credible history and large customers within the Government and Public sector, provide further international coverage with multiple continental reach (especially FJS) and lastly a possible for consideration for purchasing Fujitsu is it builds the current SPARC chips and Oracle has commited to staying in the Hardware business so I expect will want control on the production line.
Another is Cisco are a possible to buy Accenture, they work together on current engagements apparently (how cute) and I see them probably being closely matched in regards to target audience. Accenture is very expensive and has a large price tag but Cisco is one of the few orgs who are capable of buying them HP/EDS stylie. Cisco might even go for an Indian Outfit like TCS or HCL, they are into emerging markets and they probably have large outsourced operations over there already so maybe a good cheaper mix?
Whatever the next acquisition is it will raise an eyebrow, with emerging countries making the world a smaller place and price of companies being at a lower value due to the economy it is bound to happen.
In between doing some work on the Part 2 of the post series "Cloud are you ready" I thought i'd post in between with some thoughts related to Virtualisation strategy within the datacentre and how I think we need to change and knuckle down to start to get Virtualisation to do what it is supposed to do for businesses....do more with less yet provide more agility and flexibility.
Virtualisation version 1.0 in where the industry performed basic DC consolidation and sat reaping the benefits drinking Kool Aid has and still is currently the predominant phase. In V1.0 the Hypervisor provided organisations with extremely good performance factor, clever and turn key benefits to consolidate along with supportive functionality such as Vmotion, DRS, HA and excellent provisioning opportunities. It probably sounds a bit unfair to sum up current Virtualisation based on the hard work VMware and its development team has done so far, but I am sure that even VMware would admit for x86 Virtualisation to truly succeed in the datacentre and more importantly impact and become the defacto server platform within the Datacentre it needed to diversify.
This diversity was evident, they bought companies such as Dunes (Lifecycle and Stage Manager) and Akimbi for Lab Manager so hopefully this post doesn't come across like i'm saying VMware was not business ready or serious about Datacentres...they are and they got the hell of One app per server and Physical Bloated server consolidation off the ground so we could SAVE BIG BUCKS/POUNDS to invest into IT someplace else more effectively.
Enter the dragon
So enter the next phase of Virtualisation, stage 2.0 strategy which is built upon the feature rich business focused products that has evolved and grown in ecosystems of the big virtualisation companies like VMware. The stage still predominantly in the infancy phase (when say infancy stage I think my statement was backed up recently when about 5 hands out of 50 at a local London VMUG knew what Lifecycle Manager was) V2.0 builds upon and uses the solid Virtualisation 1.0 consolidation fundamentals. It uses the later end developed enablement technology in Vmware such as available orchestration tools, chargeback, capacity management tools and self service portals to streamline IT delivery and services. Further afield of Virtualisation strategy and one step higher up the technological stack is Cloud Computing (or strategy), this combines large amounts of the technology that surrounds the V2.0 Ecosystem such as web based service catalog interaction and billing, the granual chargeback tools, rich API and many more to provide this in many different service delivery methods to organisations. And whether it is a Public or Private Clouds they all have to use such technology and process automators to ensure they meet cost and service expectation.
So the question I have is do we see V2.0 strategy completely in full adoption yet within the industry and ready to evolve further? Based on feedback from events and from blogs etc I'm not entirely sure it is, Maybe I'm not asking the right audience but I do feel adoption is not large enough yet to say it is there.
Based on the fact that businesses want to do more with less and how how extremely process centric IT is today I do feel it is the right thing for IT in organisations to move to this V2.0 engine and fast. Without this automation and interaction more with our business processes we will suddenly start to hit the same issue that were hit with Physical sprawl and our Virtualised worlds will mean an untenable situation to provide justification for further investment.
Cloud computing and overall Strategy in its current lifecycle is mostly at the early adopter stage. However with the popularity of methods such as metered server usage, shrink and grow capability and per hour driven costs for workloads, based on what Businesses want from IT with the limited budgets available today means the strategy is quickly turning it into a solution which will most likely become mainstream in the next 2-3 years as a common core business strategy for delivering and supporting core services.
This Three part series of post's will hopefully provide an overview of each the component layers that formulate a cloud strategy, I will provide detail on example attributes that exist within the relevant component layers and I also provide some operational and architectural readiness advice to be able to ensure that you can exploit the cloud strategy and reap inherent benefits.
Cloud stacks
Cloud Computing is not a tool or piece of hardware you buy, it is not software you buy off of the shelf and it is not something you can buy from an SI or an Outsourcer, so to your average IT professional it has very confusing definition statement about what it actually is which can be very misleading when trying to adopt and preach the strategy to internal business leaders and potential sponsors for your initiatives.
For this series of posts I define each component that builds up the typical Cloud Computing stack. There are many views on the definition of cloud computing is, upon observation in the proprietary world it appears that Cloud Computing is being defined into the "as a Service" stack layered component approach, from which I fully agree with and embrace for simplicity. Below is a diagram (or attempt) which shows the component layers that build the Cloud stack and the relevant attributes within each Relevant "as a Service" section that we commonly know today in most areas of IT and the datacentre;
The important thing to note with the above shown Cloud stack is that each component and running attribute within is portable and naturally decoupled from each layer and can easily shrink/grow on demand according to business requirement.
For a model example of something within this stack is a typical Web application, for this app the underlying Server workloads are needed (IaaS), the Middleware to run session data and to broker info to/from a database is needed (PaaS) and the presentation layer of a Web application that feed data's from the end users (SaaS) is used.
Cloud Computing is defined as all of these services and components glued together to become a complete extensible engine for your Datacentre. To gain a more broader picture of how the Cloud and inherent stacks fit together privately and publicly check out my post on the future vision I have on operating cloud services across public and private based cloud infrastructure.
So is your current IT ready for cloud?
Adopting a service oriented view of hosting infrastructure whether Privately or Publicly means you will need to introduce some changes and amendment to how you operate IT today and how your Business processes operate. You probably think you are going to struggle to adopt a cloud strategy, however you are not the only organisation who won't initially be able to successfully transition to cloud based services with your current IT and process.
Most organisations big or small have barriers in the way for adopting new strategy, for the larger enterprises is usually a financial barrier, current outsourcer lock in, or a people political barrier and for smaller sized organisations it maybe that you have again invested in Infrastructure and cannot justify throwing away kit or you just don't realise what you can achieve from using cloud strategy. In the next sections I set the scene on some hot areas you can focus on developing and progressing to ensure you can work towards implementing a cloud strategy.
The first steps
In order to put what you have today into a "Cloud" perspective I am firstly going to talk about IaaS which is at the lowest denominator of the cloud stack. If you have Server Virtualisation in your datacentre you are certainly in a good position to start to transition into cloud strategy at the IaaS layer. If using Private Clouds you can establish yourself as having basic foundation capability for hosting decoupled servers that can move from one IaaS provider to another. If starting to deploy a private cloud you have flexibility to perform fast deployments on the fly and you also have granular visibility of metered usage to bill out to internal customers.
Because everything whether it is in a private cloud or whether hosted in a public cloud needs to be paid for in some shape or form some of the questions you need to ask yourself about your virtualisation estate being capable for an IaaS strategy are;
- How does your business pay for Infrastructure services today? - Is it on Capex, is it Recharged to business units on an Opex basis, is it on lease, do projects pay for individual requirements on an ad hoc basis,
- Do you chargeback on services based on usage and consumption? Probably answered in question One but if you currently chargeback this is a good thing, IT units see cost as an Opex which is one of the clouds benefits.
- If you do chargeback for VM's how do you do it? - Do you use a tool or do you have a rule of thumb per Virtual Machine that you use internally?
- Do you use Orchestrational capability - Do you deploy VM's through a web portal which is aligned to a Service Catalogue?
Whether you do any of the above is most likely for the majority to be a no, a lot of companies have only just started to adopt the strategy for how Infrastructure is billed and how it gets paid for using Chargeback and Opex based financing in your business. If you have implemented such processes aligning and moving to a Cloud strategy will be far much easier. The below diagram is a conceptual on benefits which can be achieved by combining Orchestration and Chargeback models within current virtualised infrastructure, Like I said on my questions above if you have this today you are at the extremely mature end of adoption for your current Infrastructure and will most certainly find moving into a Public or building your own Private cloud easier than most.
If you don't have the above process and underlying technology to support such a process flow for service requirements from your business, define your requirements from the business and then look to relevant tools and solutions which can enable them within your Infrastructure.
Once you have the "engine" and the IaaS platform that can be exploited to support a complete cloud it is then important to look at how your costs are defined for the provisioned services that your customers require and want. It is important to know that even within the Cloud world cost is based on Opex and not up front Capex for services. With the cloud the end goal is that you only pay for what workload you want, you can shrink and grow application landscapes and financial to avoid the upfront capital expenditure to procure more physical Infrastructure to support any growth and burst.
The key for moving into the start of Cloud is to ensure you have processes in place, my diagram is a conceptual vision and is by all means not going to be completely relevant to your organisation, and with that you need to engage with who is responsible in your business for budgeting and finance and ensure you align this to cloud based strategy.
Summary
Hopefully for the first post in this series I have set the scene as to what the basic components that exist within a Cloud are and what layers build up the cloud. Additionally hopefully you have learned about what the first component in cloud strategy, Infrastructure as a Service is and how you can possibly start to move into a cloud strategy a lot quicker than you may think.
In the next post I will describe what PaaS is and how I think the common OS will operate and function in the future PaaS Cloud stack. Additionally above this layer it will be interesting as I have no idea how middleware and such like works so it will certainly be an interesting read for you I'm sure!
After attending VMworld last week I was pleasantly surprised to see some of my previous blog posts almost being based what VMware are using as future strategy, things such as the use of vCloud and IO DRS were two in particular, and no I have not had an NDA for many months.
The event was about average compared to previous years, vendors were strive with Marketing in full 5th gear and the word C*oud was being used on I kid you not probably every vendor stand except the beer stands on the welcome party. Amusing as it may sound this was rather frustrating, I am a believer in cloud and the architectural principles that surround it however simply bolting on the word cloud onto your sales pitch simple doesn't cut it with me, if anything it sends me the other way.
Back to the show and on the Tier 1 vendor stands I got to spend some decent time with Dave Graham talking EMC Atmos, checked out Vmax, on the Cisco stand I saw some of the cool UCS kit (and had a laugh selling it to some random guy), spent some time struggling to find out if Layer 2 was only supported with Long Distance Vmotion (don't ask), spent moderate amounts of time asking questions on the VMware stands on new tech, and got plenty of free useless goodies that I will probably never use/wear.
Once i've had time to do some time studying sessions I attended in more detail and look at sessions I wasn't able to attend I will be most certainly writing some core material. Amongst many these will include discussion on Spring source and vCloud direction in further detail. After attending detailed sessions on vCloud I've started to understand in more detail how vCloud will work, I will endeavour to provide some technical learning material, views and opinions on this. Will provide some posts on Storage IO DRS based on what VMware.
Keep posted folks!!!
The important thing to note with the above shown Cloud stack is that each component and running attribute within is portable and naturally decoupled from each layer and can easily shrink/grow on demand according to business requirement. 