Cloud Computing

...

Economies of Scale and Pricing

The infrastructure of AWS is unprecedented, with more than 17000 cores in the EC2 platform, thus positioning AWS as the 42nd super computer[4] of the world. Amazon, by using its bargaining power could achieve large cost efficiencies through high-volume, low-cost procurement. Since AWS started in 2007, its prices have been reduced more than 19 times. New competitors like Google Cloud Engine[5], has put pressure on AWS to further reduce prices. However, customers still prefer AWS, because of the 25 odd variety of services available under one umbrella. AWS provides Database services, along with storage and computing, and at a discount price if a bundle of services are selected from AWS.

Positioning: Competition and Coopetition

With AWS, Amazon is in direct competition with Microsoft and Oracle Cloud, yet AWS also uses components from Microsoft and Oracle. The Amazon Relations Database (RDS), uses Microsoft’s SQL Server and Oracle Database Server. Amazon is one of the largest customers of Oracle and Microsoft for their database products. AWS partners with the database providers to offer value-added services to AWS customers. The customers substitute their Oracle or Microsoft Database products with AWS cloud service. Thus Amazon is in a coopetition with Microsoft and Oracle.

AWS has positioned itself as the one-stop-shop-for-all computing needs for starting a new business where technology is involved. Since Amazon itself is the biggest internal customer of AWS, it has invested wisely in AWS to strengthen its own business assets. Other players like RackSpace[6] are uncertain about new products and initiatives, and therefore do not pose a threat to AWS. Huge costs are involved in setting up a client company on AWS, and there is a cost for exit too. These costs prohibit customers from switching to another cloud service provider and AWS continues being profitable.

By introducing cloud-based music service to sell MP3 songs in 2007, Amazon was in a head-to-head competition with Apple’s iTunes. Amazon’s music content was sold at a cheap price due to the cost benefit Amazon derived from its economies of scale. Amazon Cloud Drive, also introduced around the same time runs on AWS. Kindle sold by Amazon is cheap due to the free 500GB of cloud storage that is available on the device. Amazon App Store, which was built to sell Android apps, is a popular Value Added Service to AWS.

Consumerization of Business

On-demand service provisioning: Entrepreneurs can test their business ideas with a small investment in AWS. Dropbox and Airbnb [7]are the most popular IT start-ups that successfully run on AWS.

Elasticity: Certain businesses, especially social media platform, start small but grow exponentially. For them AWS is a perfect fit, since it provides elasticity to scale up or down the business.

Inexpensive: If a business were to purchase and configure certain products like Oracle Server[8], it would carry substantial costs. Security would also be a concern for small to medium businesses. These activities are optimized and taken care by AWS for its customers, thus bringing down the total cost of ownership for its customers.

Zero or negligible IT labour costs: By using AWS, companies can continue to develop their core competency areas. Cloud Computing is enabled by data centres and virtualization technology. Virtualization creates a simulated environment that can run software just like a physical computer.

The Game in Cloud Service Provisioning

Google, Amazon, and Microsoft are offering Cloud Computing services such as Google's App Engine and Amazon's Elastic Compute Cloud (EC2) or Microsoft[9] Windows Azure. Problems arising in the ICT (Information and Communications Technology) industry, such as resource or quality of service allocation problem, pricing, and load shedding, cannot always be handled with classical optimization approaches because each player can be affected by the action of all players, not only by his own actions. In this context, game theory models and approaches allow to gain an in-depth analytical understanding of the service provisioning problem. The cloud service competition can be perceived as a game in Nash equilibrium, that is, no player can benefit by changing his/her strategy unilaterally or, in other words, every player is playing a best response to the strategy choices of his/her opponents.

Like other SaaS (Software as a Service) providers, AWS aims towards minimizing the cost of resources and cost of penalties for software failures. A challenge for service providers like AWS, is that the workload received is dynamic and fluctuates throughout the day. Flexible allocation of resources is needed at run-time. Since the underlying hardware is the same, each SaaS competes with other SaaS in the system to try to capture the maximum computing resources possible. Thus the situation corresponds to that of a Generalized Nash Equilibrium. The run-time service provisioning problem is modelled as a Generalized Nash Equilibrium Problem (GNEP), and we use non-cooperative Game Theory outcome.

Following GNE and basing price-quality competition on firm technology, providers can manage a profit despite prediction that the market should be totally commoditized. Since cloud providers use similar, if not identical, physical hardware they cannot differentiate their products on the basis of fixed costs and thus profit margins should converge to zero.

Product differentiation is achieved by benchmarking workloads across two major provider's various service levels (“virtual machine" (VM) size). We find different run-times for similarly described offerings, such as “2 virtual cores, 4GB memory." While runtime decreases for both providers as one moves to larger VMs, the price-performance trade-offs are different, which means there are different feasible price-quality combinations.

There are characteristics of monopoly in the markets where AWS operates. SEC Filings reveal that AWS is currently many times larger than the next closest competitor. Moreover, a customer of AWS faces huge switching costs, leading to potential monopolistic tendencies to keep a customer “locked in". Quality is a major condition for success of a software service firm, however offering an additional quality level does not guarantee more revenue in the case of AWS or Cloud service. When the quality level is increasing almost linearly in price and there are some customer types in the system that are highly sensitive to delay, offering additional higher quality products,