As I quoted VMware CMO Rick Jackson in my previous blog, cloud computing is a journey. You may wonder, “If so, what are the phases for that journey?”
A great question! Understanding the phases helps us to better plan our journey and have a smooth transition from phase to phase.
Bothered by SLOW Web UI to manage vSphere? Want to manage ALL your VMware vCenters, AWS, Azure, Openstack, container behind a SINGLE pane of glass? Want to search, analyze, report, visualize VMs, hosts, networks, datastores, events as easily as Google the Web? Find out more about vSearch 3.0: the search engine for all your private and public clouds.
In this blog, I will introduce 3 phases one by one: 100% virtualization, private cloud, and federated cloud. I will go over each of them and discuss the challenges in each.
1. 100% Virtualization
The concept of cloud computing is not new. You may still recall on-demand computing, autonomous computing, utility computing model, etc. These terms have come and gone, therefore people naturally have reservation on the future of cloud computing this time.
Unlike its predecessors, cloud computing is now backed by a solid foundation, which is virtualization. Once your computing resources are virtualized, you can do many things that are impossible otherwise.
The milestone for this phase is quite clear: 100%. When talking about 100%, it’s a challenge itself. The percentage is too stringent, isn’t it?! Once you get majority resources virtualized, you are good to move to the next phase.
2. Private Cloud
In this phase, the focus is to drive ROI from previous phase by sharing the resources across an enterprise. A connected and unified virtual infrastructure is the starting point, from where value added services are needed.
These services include but not limited to:
- Administration Automation. Common administrative tasks are fully automated, driven by high level policies.
- Self Service with Workflow Capabilities. Employees can request computing resources via Web based workflow management system.
- Visibility and Reporting. It helps get insights on what’s going on in the cloud, for example how resources are used with concrete data. Also make chargeback available even if there won’t be a real dollar movement from business departments to IT.
While implementing the private cloud, an enterprise can also move up in the stack from IaaS to PaaS, and even SaaS. I have written a short article on “DIY PaaS” at vSphere Java API homepage and reposted at a SpringSource blog. The key benefit of PaaS is to have all the benefits of “PaaS” but no vendor lock-in.
3. Federated Cloud
To avoid one time capital spending, enterprises would like to “borrow” into public cloud. Initial use cases will be test projects, pilot development/research projects, peak time workload balancing, etc. The common characteristics are non mission critical system with no or limited confidential data.
For big enterprises, it’s all possible to use multiple cloud service providers at the same time. This may be caused by geography, specialty, level of services in the stack, service differentiation, and etc. For example, project A may use SP X’s IaaS service; project B may use SP Y’s PaaS, and so on.
At this phase, the key challenges are:
- Security and Compliance. It’s not a new challenge – you need these even before your cloud journey. But when things are moved to outside, people naturally concern these more than before. With more experiences with the public cloud, they will be more comfortable with the external computing services.
- Federation Framework. The framework facilitates the resource unification and optimization, VM migration. Using the external services should be as easy as internal one, and as secure as internal one.
- Single Glass Pane for Management. With assets in and out the premise, IT needs unified views of every aspect. It may or may not have the resources managed as a single big pool initially, but that is the direction to go.
We’ve discussed the 3 phases in the journey of cloud computing for enterprises. There will be overlaps of different phases in reality, especially in big enterprises where IT infrastructures are distributed across different locations. That is totally OK. The important thing is to understand the progressive nature of cloud computing and plan the adoption accordingly.