Feb
03

8 Cloud Building Conditions You Need for Taking your Data Center to the Next Level



   
 

Nearly everyone down the line in IT (from service providers to enterprises) are interested in implementing cloud-based services. The attraction of cloud computing is based on the idea that it enables businesses to quickly access hosted private and public resources on-demand, decreasing the time requirements and complexities that are typically associated with physical IT infrastructure, particularly the investment, installation, configuration and deployment.

One of the biggest challenges faced by nearly all "cloud players" is the complex integration of the variety of software and hardware components from various vendors. The system that results from this integration can often be expensive to develop, and equally if not more hard to operate. This can even nullify the motives for moving to the cloud in the first place.

It is for this reason that it is important for all players, from service providers to enterprises (Large, medium and small), to fully understand some key requirements for building developing cloud management for their infrastructure.

 

Condition #1 Dependability and Availability

As models and infrastructures are focused around cloud computing, much changes with regards to how IT services are delivered and how they are used. End-users expect the cloud to be up and running on-demand 24/7 – whether we are talking about serving as an application testing ground or delivering the next big thing in multiplayer video games. IT services are delivered to/or consumed by an end-user within an organization or an outside client and there is absolutely no time for downtime. To be fully dependable and available, the data in the cloud infrastructure must always be active, no matter if a failure strikes one or any combination of its components. This level of availability is vital, since many times the architecture of a cloud-based system relies on resources that are shared across a multitude of groups that are both external and internal to each group.

Condition #2 Automated Configuration Control

As cloud infrastructure relies significantly on virtualization, it demands even more support for dynamically changing environment architectures and system configurations. This will require automated configuration management processes that focus on control components.

This shift towards virtualization technologies impacts configuration management and change management:
  • The Configuration Map becomes increasingly dynamic. This is a result of virtual machines transitioning between hosts and the frequent instance initiation and shutdowns
  • Reduced human involvement in actual deployment and maintenance actions and replacement by automated tools. The automated tools and virtual infrastructure introduce an encapsulation layer limiting visibility into the system configuration of a guest system.
  • When an affected system runs in production,  a change failure can  have a major impact on performance and availability.


This situation will demand new requirements for configuration and change control in 2011.

  • Support of dynamically changing environment architecture and configuration.
  • Automation of configuration management processes with a focus on the change control component.
  • The ability to turn sizeable amounts of change and configuration data into actionable information.
  • Tight integration with deployment automation tools


These requirements will trigger a change in vendor strategy resulting in new types of tools for the market. These tools will be an inherent part of virtualization management and cloud platforms developed by more mature vendors, like VMware, and also will come out of the innovations of smaller vendors.

Condition #3 Supports a Mixed Bag of Infrastructure: Physical, Virtual, and Cloud

Cloud systems must be able to support a mixed bag of technologies. Wait! You don't have to get  rid of your organization's existing data center (whose infrastructure took a few years to build). Whatever cloud-based solution you choose, it should take full advantage of the available solutions from a multitude of vendors, integrating with traditional IT systems whether it be hardware, software or virtualization. You will need to make sure that the cloud based solution provider not only provides a way to fit in with any open source systems but since many enterprise class organizations have developed proprietary systems, the  cloud build out must take this into account.  Your cloud solution must have support for today's leading technologies from the likes of leading application vendeors.

Condition #4 Provides for Efficient Service Management

Cloud-based systems must be able to provide administrators with a simple means for defining and measuring service levels. A service offering is a quantified set of services and applications that end users can consume through the provider — whether the cloud is private or public. Issues such as service guarantees, resource management and billing cycles are just a few. The service management offering needs to be integrated into the initial cloud offerings. This way, defined services can be efficiently created and provisioned by the end-user from the start of the cloud platform roll out.

Condition #5 Handles Agile Workloads and Large Resource Management

Cloud-based systems can be support anywhere from a few hundred people to a few thousand, and even tens of thousands at any given time. For a cloud to be responsive at these level and truly be an on-demand and expandable system, it must be aware of its workload and understand resource availability. Once a cloud system is set in operation, solutions to manage such parameters as resources and data management must be easily available to administrators. This becomes critical during peak periods. When not configured properly, it can spell potential disaster for an organization with an outage and subsequent downtime. With service level agreements at stake, and all it takes is a single outage to jeopardize IT operations, this capability is a high priority requirement.

Condition #6 Reports on Critical Performance Information from the Cloud

The success of a cloud-based deployment relies heavily on its ability to provide reports on performance and service levels.  Without reporting mechanisms about what is happening, you will face huge obstacles in meeting service levels, maintaining performance, billing and achieving compliance. Critical information must be able to be extracted from the cloud-based environment to ensure the system remains compliant, secure and performing at optimum levels and granular levels of data capture is key.

Condition #7 Integrates with Existing Management Tools

There is a common misconception that a cloud-based system can replace their existing Data Center Management tools. This could not be further from the truth. Even though the cloud is a new platform for delivering IT services, it must be able to be integrated with existing data center management tools.  Cloud computing management solutions do not replace the variety of tools present in data centers used in provisioning, customer care, billing, systems management, directory, security and much more. These tools would  need to be integrated via application programming interfaces (APIs) from your cloud solution out of the box.

Condition #8 Provides Focused Interfaces

One of the elegant sides of a cloud-based system is the fact that all the system complexities are shielded from the end-user. This helps by driving adoption and by decreasing operating costs as the majority of the management is offloaded to the end user. End users of a cloud-based service should be able to manage their own virtual data center, create and utilize their own templates, access libraries of data (graphics and text), care for their own virtual storage space. That shifts the datacenter management to the end-user, and frees the administrator. Administrator interfaces must also function with ease, providing a single window into all of the physical resources, virtual machine instances, templates, service offerings, and multiple cloud users.
blog comments powered by Disqus

Written by Alex Gutman and Martin Perlin.

Get actionable insights Now!