Cloud computing – services that are accessed directly over the Internet – is the new buzzword in the information technology world. The ‘cloud’ label comes from the fact that the computing infrastructure is not in your hands; it is located far away ‘in the clouds’, as it were. The device in your hands (personal computer, tablet computers, mobile devices, etc) is akin to an old-fashioned computer ‘terminal’ linked to a mainframe. (Readers who finished university after the 1980s will need to look this up on Wikipedia.)
Cloud computing is often called a 'game-changer', something that will dramatically change the industry, yet there is virtually no research on the economics of it. The literature that exists has primarily focused on the technical aspects of cloud computing (Ambrust et al 2009).1 Another branch of the literature focuses on the macroeconomics effects: creation of jobs, changes in output, etc (Etro 2009).
From a technical point of view, cloud computing is indeed a revolution. A 'cloud' data server farm or cluster is a collection of computer servers maintained by a cloud provider to provide computing services on a massive scale. This scale can be used both for data storage and management services as well as providing software service.Cloud computing services will be available from any device that connects to the Internet. As of 2009, 70% of Americans already used some type of cloud service – mainly web-based email.
But what about the effect of cloud computing on competition within the IT services industry? This is something we focus on in recent research (Fershtman and Gandal 2012) and discuss here.
Cloud-computing building blocks
There are essentially three aspects of cloud computing, two of which are quite related:
- IaaS (Infrastructure as a Service) – number crunching, data storage and management services (computer servers(
- SaaS (Software as a Service) – 'web-based' applications (such as Gmail or Hotmail)
- PaaS (Platform as a Service) – essentially an operating system in the cloud such as Google AppEngine and Microsoft Azure.
According to Forrester research as quoted by the Economist, the first category (IaaS) generated sales of approximately $1 billion in 2010 (Economist 2010). This category provides data storage, data management, and manipulation of large databases. Amazon is the dominant firm in this market, with a market share of 80%–90%. The second category (SaaS) generated revenues of $11.7 billion in 2010, while the third category (PaaS) generated $311 million in revenues. But these two categories are essentially part of the same market, since operating systems have virtually no stand-alone benefits and the value consumers place on operating systems increases in the number of compatible applications that run on the operating systems. Further, the dominant firms – Microsoft and Google – with (virtual) operating systems in this market provide most of the important in-house applications available (ie, office suites like Microsoft Office and Google Docs).
The revenues of the PaaS/SaaS market are expected to grow to $52 billion in 2020, while the revenues of the IaaS market will only grow to $4 billion by 2020 (see The Economist 2010). This relatively small increase in revenues in the IaaS market is largely due to the fact that the cost of computer hardware (infrastructure) will continue to drop. This trend is similar to what happened in the personal computer market. Like the personal computer market, the IaaS market is characterised by relatively homogenous goods; hence it is likely that it will be subject to intense price competition and low profitability. In the PaaS/SaaS market, on the other hand, the products are potentially quite heterogeneous and we expect greater consumer loyalty.
- Low-cost, largescale data storage and data management in remote servers will make up less than 10% of the cloud computing market.
- More than 90% of the cloud computing revenue will involve virtual operating systems and applications software services.
Current state of competitionin PaaS/IaaS software services in cloud
Two firms, Microsoft and Google, already have well-established platforms in the cloud: Microsoft Azure and Google AppEngine.
Microsoft and Google also provide software applications for their cloud platforms.
- In the case of Microsoft, the in-house applications for the on-premise Windows operating system are offered online via Microsoft Azure as well.
These applications include Microsoft Office. The on-premise operating system (Windows) is compatible with the 'cloud' operating system (Azure). This means that all office software files (word processors, spreadsheets, and presentation software) created 'on-premises' can be edited on the cloud as well, and vice versa. Edited files are automatically synced between the cloud and the on-premise environment.
- Google in-house applications include 'Google Docs', an online version of an office suite that consists of a word processor, spreadsheet, and presentations software program that works on the Google AppEngine platform.
Google Docs can be accessed by all users with a Gmail account. Users can create the documents, spreadsheets, or presentations within the application or it can be imported from other formats and converted to 'Google Docs' format. The documents are then saved to Google's servers, but they can also be saved to a user's computer. The service runs on most web browsers. Google Docs 'documents' can be shared – and they can be viewed and edited by multiple users in real-time simultaneously.
Two key economic issues
There are several important 'economic' aspects of cloud computing that will affect the competition in the SaaS/PaaS (platform) market. Arguably, the two economic aspects that will have the most profound implications for the development of cloud computing are:
- Changes in the strength of network effects, and
- The organisational model that emerges, ie, whether cloud computing will retain its current 'conduit' and vertically integrated structure or whether it will develop into a two-sided market.
Network effects in cloud computing
Since application software programs are typically compatible with a specific operating system, network effects played a major role in determining the type of market equilibrium. In particular there were very strong network effects in operating systems in on-premise markets. Consumers preferred to use operating systems that offered a large variety of application software (indirect network effect), while software developers preferred to develop software for operating system with many users. This often led to a setting where a single firm won the battle among standards.
The transition to the cloud will likely weaken the network effect for operating systems in the cloud environment. This is because:
- Data in application software programs are more likely to be portable across operating systems.
- Platform switching costs for developers moving from one cloud to another (say from Microsoft Azure to Google Application Engine) are about the same as desktop software (Windows to Apple). At the consumer level, however, there is little or no switching cost in the case of cloud-based applications. The switching costs for changing platforms are incurred only at the server level, and the consumer may not know (and definitely does not care) that the underlying operating system has been changed. Because applications programs are accessed directly via the browser, 'compatibility' among software applications written for different operating systems will exist in the cloud to a much greater extent than in the online markets.
Weaker indirect network effects in virtual operating systems make it more likely that multiple platforms can exist in equilibrium in the cloud. Hence de facto standardisation on one platform seems less likely in competition among cloud platforms than in the case of competition among on-premise platforms.
Organisation and evolution of cloud markets
How might cloud computing markets be organised? Will cloud computing evolve into a two-sided market, or will it remain a vertically integrated system? Currently, the two main platform owners (Microsoft and Google) supply the critical complementary software (email service, office productivity suites) for their platforms in order to attract consumers.
At this stage, cloud computing is similar to cable television service in many respects. Like the cables that bring content into a consumer's home, the cloud is also a conduit. Further, in the cable industry, there is a great deal of vertical integration between the owner of the conduit and the content providers.
An alternative structure is when cloud platform providers only offer the infrastructure, ie, like a shopping mall. This form of organisation is indeed a two-sided market. In such a setting, there is a platform owner who may charge independent software developers fees in order to be able to provide their service via the platform owner's proprietary cloud. While the fee may have a fixed component, it also may depend on the number of consumer subscribers who use the cloud. In this case, pricing has all the standard issues associated with two-sided markets. (For more on two-sided markets, see Armstrong 2006, Rochet and Tirole 2006, and Rysman 2009.)
The structure that will likely emerge is dependent to a great extent on whether third-party sources of complementary products become essential for the success of the platform. In settings in which third-party provision is important, firms that do not open their platforms to third-party providers and create incentives for these firms to provide software applications for their platform often fail when competing in platform completion. Sony, for example, lost the Beta/VHS format war because of the dearth of movies available for its platform.
Additionally, different pricing models may be employed in the cloud. In an on-premise setting, a user had to buy (license) the software in order to use it. The payment was independent of usage and the licence was typically valid as long as the consumer used the software on his PC. This will likely change when we move to the cloud ecosystem, where users may rent 'cloud' software, rather than purchase it.
Other important economic issues
Compatibility: An additional key issue involves compatibility, in particular whether it is feasible for entrants to achieve one-way compatibility with an established standard. This strategy is especially relevant in the cloud ecosystem because compatibility is easier to achieve and less costly. Hence, entry is likely to be easier in the cloud ecosystem than in the on-premises market. ‘Glide' for example, is an online platform that offers email services, office suite software, and other software services. Glide's applications software products are compatible with the dominant platforms: its software runs on all three of the major desktop computing platforms: Windows, Linux, and Apple (Macintosh). Additionally, Glide software is also compatible with virtually all 'smartphone' platforms in order to attract users who access the Internet from cellular phones. In the case of Glide, compatibility across desktop and mobile platforms means that documents edited online will be updated on the desktop (or mobile) and vice versa. By ensuring compatibility with the dominant platforms, Glide increases its chances of being able to compete despite a much smaller installed base. (See Rosenblatt 2009.)
Cyber security: There is a nascent literature at the ‘intersection’ of computer science/engineering and economics on cyber security. Much of the work in the field has been undertaken by computer scientists and has led to some important insights into why information systems have become so insecure (Anderson and Moore 2006). Contributions by economists have primarily focused on (i) the lack of incentives for individuals or network operators to take adequate security precautions – see Varian (2004) and Camp and Wolfram (2004) among others – and (ii) the incentives for firms to disclose information about vulnerabilities (Choi et al 2010.)
There is now a consensus in the literature that improving Internet security involves more than just finding technical solutions. Indeed, the Slammer, Blaster, and Sobig.F viruses exploited vulnerabilities even though security updates had been released. That is, although the updates were widely available, relatively few users had applied them. This clearly illustrates that technical solutions alone cannot solve cyber security problems.
As the paradigm shifts so that more SaaS/PaaS services are provided online rather than on-premise, the exposure to such viruses and thus the potential for damage is even greater. Recently, engineers and computer scientists have expressed concerns of an increase in security breaches (Molnar and Schectter 2010) as the industry switches from in-house computing to services offered by ‘public’ (ie, independent) cloud providers. An important question is how the industrial organisation of cloud software markets affects the incentives of cloud providers to implement effective security.
Cloud computing is still in a very formative stage, and it is not possible to foresee how platform competition will develop. In our research, we examine how key factors might affect the development of platform competition in cloud-based computing. Our goal is to raise economic issues that will likely affect the development of SaaS/PaaS services in cloud computing.
Armbrust, M, A Fox, R Griffith, A Joseph, R Katz, A Konwinski, G Lee, D Patterson, A Rabkin, I Stoica, and M Zaharia (2009), “Above the Clouds: A Berkeley View of Cloud Computing”, EECS Department University of California, Berkeley, Technical Report No. UCB/EECS-2009-28.
Anderson, R and T Moore (2006), “The Economics of Information Security”, Science, 314:610-613.
Armstrong, M (2006), “Competition in Two-Sided Markets”, The RAND Journal of Economics, 37(3):668-691.
Camp, L and C Wolfram (2004), “Pricing Security”, in LJ Camp and S Lewis (eds.), Economics of Information Security, 12, Advances in Information Security, Springer-Kluwer.
Choi, J, C Fershtman, and N Gandal (2010), “Network Security: Vulnerabilities and Disclosure Policy”, Journal of Industrial Economics, 58:868-894.
Etro, F (2009), “The Economic Impact of Cloud Computing on Business Creation, Employment and Output in Europe”, Review of Business and Economics, 54(2):179-208.
Molnar, D and S Schechter (2010), “Self Hosting vs. Cloud Hosting: Accounting for the security impact of hosting in the Cloud”, Workshop on the Economics of Information Security (WEIS), Harvard University.
Rochet, J and J Tirole (2006), “Two-Sided Markets: A Progress Report”, The RAND Journal of Economics, 35(3):645-667.
Rysman, M (2009), "The Economics of Two-Sided Markets, Journal of Economic Perspectives, 23:125-143.
1 This is an excellent primer on the development of cloud computing and issues related to data storage management, computation and related services.