Cloud potential to scale is highly dependent upon a smart network infrastructure
Cloud enterprise architecture is the new apps-on-tap, utility computing, central mainframe/thin client orthodoxy (care to recall the time-share computing days … hey, FORTRAN is even still around!).
What’s new is old; except in the case of enterprise cloud, the net-new requirement is optimizing the underlying network infrastructure to deliver reliable, secure, compliant scale and speed for on-demand provisioning of business-critical applications, with minimal human (read: professional consulting) intervention.
At odds with enterprise cloud is the traditional outsourcing services model. Yes, you may end up paying less over time but your deployment, systems integration and app development/provisioning time-to-market is no greater, and in fact, often turns out to be more elongated. What’s more, the delay in deployment and absorption risks loss of competitive and revenue advantage.
Inherent in this dilemma is what Rob Whiteley of Forrester terms the “automation imperative.” Whiteley’s Law says that the greater demand for the business to evolve its IT model, i.e., from client/service to cloud, the greater the IT skills gap grows, thus the imperative to automate as many execution elements of the network that delivers the necessary capacity and intelligence to make cloud work.
Another challenge to delivering the quality of service expectations of cloud enterprise architecture is the inherent nature of scale across thousands of users and multiple geographies; for example, a Japanese multi-national conglomerate was able to leapfrog the competition with a new on-line photo-sharing service built in a cloud-based development and test environment residing totally in the network. Implementation of the environment across multiple geographies and thousands of users allowed the company to experience a five-fold decrease in conventional systems integration and development costs.
In order to solve the automation imperative and scale dependencies, critical to the success of cloud enterprise architecture, is to first employ a robust network lifecycle management initiative. Network lifecycle management can enable the enterprise, for example, to gain trusted visibility to real-time information about what is installed in the network and how devices are connected, even to the granular degree of the exact device feature set and OS versions, and network topologies configured inefficiently, and thus which are prime candidates for evolving to cloud.
An automated view of the network-level device inventory can also aid the organization in determining regulatory compliance standards, security policy adherence and even network performance metrics in terms of application delivery and business process optimization. For example, greater transparency to network performance systems that are smart-enabled can send automated intelligence data feeds to network managers if certain components in the network are not functioning properly, empowering administrators to quickly and proactively prevent potential IP network issues and deliver a more predictive user experience.
Organizations will realize even more value for their cloud implementations by also examining the return-on-investment of the “operate” phase of the initiative, which often is several times the cost of the capital investment. A key way to reduce operate costs is the minimization of bandwidth overcapacity. Traditionally, performance problems are often addressed by throwing more bandwidth at the problem. If the problem is on the application or configuration side, however, a costly increase in bandwidth will not necessarily solve the root cause and inadvertently increase long terms costs.
In cloud enterprise architecture, business processes and applications are virtualized in the IP network, and as such, network performance and availability are linked to profitability and shareholder value. A comprehensive network infrastructure transformation campaign can lay the groundwork for a successful cloud implementation by determining at the continuity phase how to demonstrably increase the knowledge of the network control points, reduce potential security and compliance impacts, and accurately balance performance versus cost tradeoffs.
BTW please join us Wed., March 30th at 9AM Pacific for a special event exploring new, innovative Data Center Fabric solutions <http://bit.ly/hsHuUx> .
For more information about Cisco’s cloud strategy, visit http://www.cisco.com/go/cloud.