Changing the delivery of IT

Tony Bishop

Subscribe to Tony Bishop: eMailAlertsEmail Alerts
Get Tony Bishop: homepageHomepage mobileMobile rssRSS facebookFacebook twitterTwitter linkedinLinkedIn

Related Topics: Virtualization Magazine, SOA & WOA Magazine, ERP Journal on Ulitzer

Virtualization: Article

The Role of the Network in Virtualization

From Silos to a Service-Oriented Infrastructure

Evolving to a Service-Oriented Infrastructure
Virtualization is just one part of a multi-phase journey toward full automation and a policy-based, adaptive infrastructure. Organizations typically begin with some level of storage consolidation, data center consolidation, server standardization, and branch consolidation, all of which immediately contributes to improved asset utilization and efficiency. Virtualization has moved beyond the early adopter stage, especially in server virtualization, where companies are beginning to incorporate hypervisor technology into production systems and are experimenting with ways to use this technology to move applications transparently from one physical environment to another. According to Gartner, more than half the companies participating in 2007 IT infrastructure surveys were well along the infrastructure consolidation path. IDC concurs, predicting that virtualization will become a mainstream technology over the next few years.

At the same time, companies are recognizing the need for more advanced tools and processes to ensure that virtualization technologies can be implemented in production environments without increasing operational complexity. While virtualization brings many benefits, it also brings the IT staff new operational challenges. Service-orchestration solutions help address some of the challenges associated with maintaining separate physical and virtualization operational domains, while also easing the progression toward a more automated service-oriented infrastructure (SOI). (Figure 1)

Virtualization: Before and After
Let's examine a typical enterprise before it implements virtualization, where its choice is to share all or nothing. If a single physical application switch is used, applications must compete for resources, changes to one application can affect the others, and the device configuration is overly complex. Adding more physical switches creates an inefficient isolation of applications and results in device sprawl, under-utilized resources, and complexity in upgrading.

Conversely, with a virtualized architecture, abstraction and partitioning allow one physical switch to provide multiple virtual contexts, enabling isolated, secure applications with guaranteed resources and role-based access. This results in dramatic reductions in provisioning cycles, operating expenses (OpEx), and power requirements.

We can compare the evolution of data networking and the Internet with the new virtualized IT. In early data networking and wide-area communications, the problem was heterogeneous "network silos" comprising disparate transport and operating systems. The various protocols included ATM, FDDI, Ethernet, Token Ring, AppleTalk, Banyan Vines, Novell Netware, and DECnet that were costly, complex, and difficult to scale. Internet 1.0 delivered "information over IP" and provided a pervasive neutral medium with standard protocols for communicating globally and sharing information.

Similarly, within the enterprise, the problem is heterogeneous "infrastructure silos" - storage networks, applications, network equipment, servers, and interconnection technologies typically segmented by department or physical location - that are costly, complex, and difficult to scale. The solution is Internet 2.0, or "infrastructure over IP," providing a pervasive neutral medium, standard protocols, and built-in advanced capabilities for collaborating, sharing resources, and lowering costs.

Benefits of Virtualization
A significant advantage that comes with virtualization is reduced power consumption. Consider a typical enterprise that requires 10 Gbps load balancing, 20 Gbps firewall protection, 10 virtual contexts, and high availability. Using virtualized integrated modules that use approximately 11kW per hour, the enterprise will see an 85% reduction in power, providing more than $90,000 in OpEx savings over three years. Rack space is also reduced by nearly 30 rack units, with additional savings gained from reduced cabling, port consumption, and support costs.

Provisioning is also greatly simplified. To add one server to a Web farm in an existing silo'd infrastructure, multiple steps involving multiple departments and skill sets are required as illustrated in Figure 2. With any delay across the multiple layers of coordination, this simple expansion can take up to 90 days. New service turn-ups can stretch beyond 180 days. The "serial workflow" problem must be eliminated to streamline new service provisioning.

Automating infrastructure provisioning via the network uses virtual service templates and physical "pods" of servers, storage, and network equipment to accomplish the steps previously done by physical organizations and people. See Figure 3.

End-to-end service virtualization via the network pays off in several other ways besides reduced power consumption. Improved utilization, total cost of ownership (TCO), time savings, and flexibility are other benefits. For example, let's assume a traditional monolithic data center that experiences 25% year-over-year growth and 20% efficiency. In the second year of use, it will reach its power and capacity limits, suffer low asset utilization and escalating costs due to high operational overhead, and provide a low business value. Building a new data center can cost up to $250 million, plus an additional $25 million a year in operating expenses.

What happens when virtualization is deployed? Using metrics that Cisco discovered during its own data center redesign, utilization increases to 60% and lost capacity is regained. With virtualization, utilization rates increase dramatically, extending the life of the existing data center an average of four-and-a-half years. The average cost to maintain a new data center is $40 million a year, so over that period of extended life, savings will grow to $180 million. Other benefits become apparent as well, including the deferral of purchasing new capacity, improved asset utilization, better power efficiency, and lower TCO.

Phased Implementation
An enterprise committed to moving to a service-oriented network infrastructure will achieve the greatest success by following a phased approach that allows for gradual consolidation and virtualization, leading to the ultimate goal of automation. Figure 4 outlines the key elements of this phased approach.

  • Consolidation Phase: Starting with isolated resource islands and disparate networks, the first evolutionary step is to consolidate the computing and storage islands into enterprise-wide networks. One way is to consolidate data center resources into fewer physical locations. Another is to consolidate storage resources into single SANs, using virtual SANs (VSANs) to allow the consolidation of SAN islands onto a single fabric while ensuring scalability and security.
  • Virtualization Phase: Virtualization allows computing, network, and storage resources to be dynamically partitioned, provisioned, and assigned with ease to different applications. Logical server partitioning, blade server deployments, and application-aware load-balancing services are all part of this phase. Storage resources are pooled into one or several shared storage resource pools. Server processing resources are pooled into one or several shared processing resource pools. This virtualization improves agility and makes it easier for the data center to keep up with changing business conditions. Resource virtualization requires the support of intelligent networks that are aware of applications and can respond to changing conditions to optimize the performance of each application. Content switching and application-oriented networking are examples of application integration.
  • Automation Phase: The final step is flexible service automation, which allows an intelligent network fabric to detect and respond rapidly and automatically to the applications' changing needs, and to provision processing and storage and security resources as needed. Automated service provisioning, automated security responses, and self-healing systems are the cornerstones of this phase. Automating these processes not only speeds an enterprise's response to new customer needs, but also significantly lowers TCO by lessening the need for manual intervention and reduces the under-utilization of allocated processing and storage resources. It also improves reliability by reducing complexity and the need for human intervention.

More Stories By Peter Linkin

Peter Linkin brings over 16 years of experience in the Application Integration and Middleware business to his current responsibility as a senior Marketing Manager at Cisco for Data Center Solutions. Prior to joining Cisco he was Senior Director of WW Product Marketing at BEA Systems, where he helped lead their entry into the SOA and Application Platform spaces, and previously held senior product management and marketing responsibility for pioneering middleware technologies at Vitria, Lucent and Hewlett-Packard. He has been a frequent speaker at industry conferences and contributor to specialist publications. Mr. Linkin holds degrees in Physics and Business Economics from the UK.

Comments (0)

Share your thoughts on this story.

Add your comment
You must be signed in to add a comment. Sign-in | Register

In accordance with our Comment Policy, we encourage comments that are on topic, relevant and to-the-point. We will remove comments that include profanity, personal attacks, racial slurs, threats of violence, or other inappropriate material that violates our Terms and Conditions, and will block users who make repeated violations. We ask all readers to expect diversity of opinion and to treat one another with dignity and respect.