Carving out a cloud computing service business case

Michael Kennedy, Network Strategy PartnersCloud computing services are emerging as new value-added service offerings for network-based service providers. They create business opportunities for network-based service providers by allowing them to use network intelligence-routing and IP policy management-to tradeoff data processing, storage, and network resources to reduce the Total Cost of Ownership (TCO) of cloud computing service delivery.

Technological progress on several fronts is creating opportunities for service providers to profitably offer cloud computing services. Server and storage virtualization breaks the linkage of applications to specific servers and storage systems as well as to single hosting sites or data centers. Progress in routing, DPI (Deep Packet Inspection), flow-based routing, and IP policy management enables policy-based routing of traffic across multiple IP/Ethernet networks. The result is that subscribers to cloud computing services can be dynamically assigned to the lowest cost computing and storage resources using economic optimization algorithms that tradeoff the costs of storage, processing and transport that vary by time-of-day and location. This approach to minimizing the cost of cloud computing services is analogous to the electric power grid where regional Independent Systems Operators (ISOs) minimize the cost of electric service by using the power grid to bring power generators online according to each generator's unit cost.

The business cases for two cloud computing service offerings--cloud computing and video-on-demand (VoD) service delivery--illustrate the revenue opportunities for network-based service providers. In the cloud computing business case, the transport network is used to reduce the peak capacity of the data centers in the cloud--this significantly reduces total data center CapEx and associated OpEx. The video-on-demand service delivery business case rests upon the use of "intelligent networking" to reduce transport costs for VoD delivery by connecting each video subscriber to the closest video server or cache that contains the requested video. This business case makes sense for both VoD offered as part of Triple Play and Over-The-Top (OTT) Internet video service offerings.

The cloud computing opportunity can be offered to a retail cloud computing service provider such as Amazon, an enterprise operating its own data centers, or as a cost reduction initiative for a service provider that owns and operates its own transport network and data centers. There is a lot of economic leverage in shifting computing demand from peak to off-peak time periods. Intelligent network transportFor example, Amazon's Elastic Compute Cloud (EC2) is a web service that provides resizable compute capacity in the cloud. The service is metered in terms of instances that are defined by the amount of memory, processing, and storage offered. A small on-demand instance running Windows is priced at 12 cents per hour. A spot instance runs on unused capacity where bids are made to clear the unused capacity. The nominal price for a small spot instance is five cents per hour. The nominal spot instance price is tied to the variable cost of operating the data center since Amazon would lose money and violate antitrust laws if services were offered below variable/marginal costs.  In this example, therefore, seven cents or 58 percent of total cost is associated with providing peak capacity.  Intelligent network transport can be used to redistribute service demand from data centers operating at peak capacity to those with excess capacity, thus, reducing the size of each data center and the total cost of peak capacity. This opportunity is greatest when a global cloud is employed and transaction processing requirements are heavily concentrated during the business day--for example stock trading in New York, Hong Kong, and London. We performed a total cost of ownership analysis that showed that under these conditions a 36 percent cost reduction could be achieved.

The video-on-demand opportunity considers a content distribution network where video content is stored at a national content distribution site, the super head-end of a metro network and at several distributed caches located within the metro network. In this case intelligent networking is used to identify and route each VoD subscriber's request for a video to the nearest content source rather than routing every request back to the content stored at the super head-end or national site. Our total cost of ownership analysis showed that savings are as high as 55 percent. Primary sources of savings are reduction in the cost of the aggregation network and in Internet peering expense.

The business case examples show that cloud computing creates attractive opportunities for network-based service providers. However, the analysis also shows that the cloud computing opportunity is not a simplistic one of substituting network resources for computing and storage resources. In fact the video-on-demand opportunity creates value by reducing the size of both the aggregation and core networks.

Michael Kennedy is a regular FierceTelecom columnist and is the co-founder and Managing Partner of Network Strategy Partners, LLC (NSP) consultants to the networking industry. He can be reached at [email protected]