The next-generation dispersed data center architecture
With the right infrastructure, service providers can deliver differentiated and profitable commercial cloud services from their data centers independently or as a complement to their internal private cloud. It’s a logical step. Cloud services are a natural extension of the virtual private network (VPN) and hosting services service providers offer today. They’re also a huge business opportunity.
To take advantage of the cloud services opportunity, service providers are transforming legacy standalone data centers into an agile, intelligent, integrated next-generation data center infrastructure. This virtualized, policy-driven, logically consolidated but physically dispersed infrastructure combines service providers’ greatest assets — their networks and expertise — with virtualization and dynamic management. It supports new service models at the scale necessary to meet demand. And it increases operational efficiency to ensure profitability.
Service providers’ greatest advantage: The network
The network is the backbone of the next-generation infrastructure. It allows service providers to interconnect geographically dispersed data centers and take advantage of data center virtualization. A network-based data center strategy allows service providers to:
- Lower total cost of ownership (TCO): Designing the data center from the network up reduces capital and operating costs. It consolidates expensive computing resources to significantly lower capital costs. Computing resources are better utilized and have a longer life expectancy. Dynamic management across the IT stack unifies and simplifies the back office. Power and cooling costs drop.
- Increase service agility: Service providers can dynamically respond to changing business demands. New services can be quickly provisioned from a shared pool of resources on an as-needed basis. Ideally, this leads to policy-driven resource allocation, which further streamlines operational procedures.
- Improve service quality: Service providers can use the reach and scale of the network to efficiently distribute cloud services across the infrastructure. This allows them to offer bandwidth and latencycharacteristics that are superior to those of non-telco cloud services providers.
- Offer better service level agreements (SLAs). Service providers can deliver the SLAs their most demanding customers require and expect. They can also combine IT and network performance into a single, unified SLA. This allows service providers to deliver higher levels of application assurance and reporting than are available today.
Virtualization and the data center WAN
Virtualization is a key catalyst for cloud computing. It’s also an important enabler for cloud-ready, standardized and intelligent data centers.
Virtualization technologies let service providers host many customers on a single server or storage array. This allows them to create multi-tenancy models in their data centers. Multi-tenancy models help service providers ensure the economies of scale necessary to deliver cost-competitive, yet profitable, cloud services. They improve asset utilization and productivity while lowering costs and energy consumption.
While beneficial, server virtualization is primarily tactical in nature. To gain strategic advantages, service providers are extending virtualization to the data center WAN. A network fabric that interconnects geographically dispersed data centers virtualizes the entire data center infrastructure. It consolidates data centers across the WAN to bring service providers the benefits of a shared infrastructure.
A virtualized data center infrastructure offers all of the efficiencies of server virtualization. It also gives service providers a truly agile cloud services environment. For example, they can:
- Host a customer’s service in the optimal data center and on the optimal platform to meet customer requirements and ensure fast application response times. This increases the service provider’s ability to meet SLAs and drive profitability.
- Deploy widespread business continuity and disaster recovery mechanisms. This increases application availability and allows customers to access applications even if a site fails.
- Balance workload demands by moving large workloads among geographically distributed data centers. This avoids demand hot spots and allows service providers to fully utilize the capacity available across the network.
Distributed data centers lower costs
While a distributed data center architecture is not right for every service provider, Alcatel-Lucent Bell Labs modeling results confirm it is cost effective. In one scenario, Bell Labs examined the relationship between bandwidth consumed per-subscriber for a data center-hosted application and the cost to deliver it. The model shows that the cost of bandwidth per subscriber eventually exceeds the cost of operating the distributed data center.
As illustrated in the upper chart in Figure 1, the cost per application increases linearly for services hosted in a centralized data center. It remains relatively stable for applications hosted in a distributed data center.
Building on network strength
As service providers transform and virtualize their infrastructure to create the next-generation data center, their network is a crucial asset. It allows them to manage large numbers of highly virtualized resources, in multiple locations, as a single source. And it provides the foundation for a truly on-demand logically centralized, physically dispersed data center that delivers computing and application resources anywhere, at any moment.
Service providers that put their network at the center of their cloud services strategy are well positioned to deliver cloud services that take advantage of their real-estate footprint, brand, culture of customer experience management and access networks.
To contact the authors or request additional information, please send email to firstname.lastname@example.org.