Cloud computing has risen to the top of IT’s discussion list, with the focus primarily on public clouds, out of the high-profile side of public-cloud vendors and their smooth accessibility. But when turning to implementation, many IT organizations are not ready to rely on public clouds out of concern for security, storage of sensitive data outside the corporate firewall and a general lack of control on the part of IT. It presents an opportunity to tremendously increase the agility of IT, allowing the IT organization to rapidly and efficiently respond to changing business requirements, while still maintaining control. However, IT execs are now coming to understand the challenges involved in making the transition to a new cloud infrastructure, where provisioning it is uncharted territory. We explore the key technical challenges that lay ahead for even the most well planned cloud implementations, and where efforts should be focused in order to […]

Cloud computing has risen to the top of IT’s discussion list, with the focus primarily on public clouds, out of the high-profile side of public-cloud vendors and their smooth accessibility. But when turning to implementation, many IT organizations are not ready to rely on public clouds out of concern for security, storage of sensitive data outside the corporate firewall and a general lack of control on the part of IT.
It presents an opportunity to tremendously increase the agility of IT, allowing the IT organization to rapidly and efficiently respond to changing business requirements, while still maintaining control.
However, IT execs are now coming to understand the challenges involved in making the transition to a new cloud infrastructure, where provisioning it is uncharted territory.
We explore the key technical challenges that lay ahead for even the most well planned cloud implementations, and where efforts should be focused in order to realize the promise of private cloud.
Defination:-
For the point of this discussion, we define private cloud as a deployment model where the cloud infrastructure is operated solely for an organization. It could be managed by the organization or a third party and could exist on-premise or off-premise.
Promises 100% Automation
To implement the enterprise cloud, today’s organizations can build a platform that enables a cloud management loop through almost any combination of free open source components, mid-range cloud management tools and commercial heavy weight system management solutions. However, this is based on the assumption of complete, 100% reliable automation which has been promoted by the vendors, and not necessarily a reflection of reality.
Painful Reality
Private cloud introduces new management challenges to IT operations.
In a very dynamic and complex environment, issues come up. This means that certain infrastructure parameters need to be adjusted and often changed manually on the servers experiencing issues. Further complicating the matter comes with updates to the deployment automation platform, adding greater room for discrepancy. It becomes very difficult to track manual changes which are frequently made at a very granular level to individual parameters. These become moving targets in the dark.
When automation assets are modified, it becomes very difficult to correlate actual changes in the environment, causing environmental drift. This results in configuration inconsistencies, longer issue resolution time and other risks.
Rollout and rollback – is this a viable approach in large environments?
Deployment automation makes the cloud infrastructure flexible and efficient, yet a major challenge for managing configuration. The reason is that in large environments the management of configuration is based on approaches like rollout and rollback of the entire server images.
Numerous organizations are leveraging cloud to rapidly scale up and down particular business systems while at the same time continuously running these systems. The Cloud’s dynamic nature, and limited visibility into the actual environment configuration makes the upgrade and roll back of the entire business service a challenge. The rollback of a system based on virtual images requires that the entire setup be updated, and then synchronized.
This introduces unnecessary operational risks. Since existing working virtual servers are retired, and new ones are rolled out, the replacement servers (based on stored images) include the changes. Many companies, operating in the cloud, are now upgrading existing servers – withoutretiring servers. This leaves servers still active with inaccurate configuration information.
The Rollback Challenge
The nature of systems that are implemented in cloud platforms is that they are subject to frequent changes, fixes, and improvements. This continues until a problem is found. Then the administrator can just roll back to the state prior to the faulty change. Many problems are only discovered after the passage of a good amount of time following release to production, and the more time that passes, the harder it is to roll back. Users get accustomed to new features, making it harder to remove the features. Customer and Business data is amassed in updated schemas, making it difficult to roll the database back.
Rollout Difficulties
Private cloud is changed and updated through several different deployment rollout models, yet each has a particular shortcoming.
A� Golden Image Deployment: The disadvantage of Golden images is that this is just combining machines ‘AS IS’. The image catalog likely will only contain the commonly used images and lack the less common combinations of components that might actually be required for specific user purposes.
A� Template-based Deployment: By following a template, a fully configured cloud instance can be deployed to any of a number of cloud environments, however this means building and configuring templates that may miss ad hoc changes that had taken place during provisioning.
A� Script-based Deployment: By using a script to build the image, you can ensure that the rollout has the latest version of necessary components, however an operator will need to configure these scripts to ensure that the rollout will be error-free – possibility introducing bugs to the script.
A� Deployment Combinations: While no single way to efficiently deploy in a private cloud environment. Any of these approaches can be mixed and combined to roll out changes in the cloud, however all of the described problems can appear in this combination approach too.
Since changes occur dynamically, quickly and often at a fine level of detail, configuration management and change management tools need to be able to stay on top of the different states of the private cloud based servers, in order to know what changed and the impact. To facilitate these deployment approaches and successfully operate in private clouds, it is critical to manage the actual environment configuration and the changes that take place in the environment. To fully realize the capabilities and promise of private cloud in the enterprise, organizations need to implement a solution that can also identify changes in near real-time, at a comprehensive and detailed level. So if your IT organization is not willing to make this investment, then when the data center is transitioned to a private cloud, it will not be a platform that exhibits agile provisioning, elasticity and lower costs per application – the actual promise of the private cloud.
Closing the Private Cloud Management Loop
The gap that leaves the private cloud management loop wide open comes from how frequently private cloud offerings ignore change and configuration management challenges, resulting in not providing the new capabilities required to support environment agility while not losing control.
To close this loop, change and configuration monitoring must be automated, implementing analytics on top of the amassed information and integrating the new process for paying attention to change and configuration in cloud management – starting at the setup phase, through migration of existing systems and to ongoing operations for private cloud.