Software as a service

Centralized hosting of business applications dates back to the 1960s. Starting in that decade, IBM and other mainframe computer providers conducted a service bureau business, often referred to as time-sharing or utility computing. Such services included offering computing power and database storage to banks and other large organizations from their worldwide data centers.[11]

The expansion of the Internet during the 1990s brought about a new class of centralized computing, called application service providers (ASP). ASPs provided businesses with the service of hosting and managing specialized business applications, to reduce costs through central administration and the provider's specialization in a particular business application. Two of the largest ASPs were USI, which was headquartered in the Washington, DC area, and Futurelink Corporation, headquartered in Irvine, California.[12]

Software as a service essentially extends the idea of the ASP model. The term software as a service (SaaS), however, is commonly used in more specific settings:

• While most initial ASPs focused on managing and hosting third-party independent software vendors' software, as of 2012 SaaS vendors typically develop and manage their own software.[13]
• Whereas many initial ASPs offered more traditional client-server applications, which require the installation of software on users' personal computers, later implementations can be Web applications that only require a web browser to use.[14]
• Whereas the software architecture used by most initial ASPs mandated maintaining a separate instance of the application for each business, as of 2012 SaaS services can utilize a multi-tenant architecture, in which the application serves multiple businesses and users, and partitions its data accordingly.

The acronym first appeared in the goods and services description of a USPTO trademark, filed on September 23, 1985.[15] DbaaS (Database as a service) has emerged as a sub-variety of SaaS[16] and is a type of cloud database.

Microsoft referred to SaaS as "software plus services" for a few years.[17]

The cloud (or SaaS) model has no physical need for indirect distribution because it is not distributed physically and is deployed almost instantaneously, thereby negating the need for traditional partners and middlemen. Unlike traditional software, which is conventionally sold as a perpetual license with an up-front cost (and an optional ongoing support fee), SaaS providers generally price applications using a subscription fee, most commonly a monthly fee or an annual fee.[18] Consequently, the initial setup cost for SaaS is typically lower than the equivalent enterprise software. SaaS vendors typically price their applications based on some usage parameters, such as the number of users using the application. However, because in a SaaS environment customers' data reside with the SaaS vendor, opportunities also exist to charge per transaction, event, or other units of value, such as the number of processors required.[19]

The relatively low cost for user provisioning (i.e., setting up a new customer) in a multi-tenant environment enables some SaaS vendors to offer applications using the freemium model.[19] In this model, a free service is made available with limited functionality or scope, and fees are charged for enhanced functionality or larger scope.[19]

A key driver of SaaS growth is SaaS vendors' ability to provide a price that is competitive with on-premises software. This is consistent with the traditional rationale for outsourcing IT systems, which involves applying economies of scale to application operation, i.e., an outside service provider may be able to offer better, cheaper, more reliable applications.[20]

Most SaaS providers offer a multi-tenant architecture. With this model, a single version of the application, with a single configuration (hardware, network, operating system), is used for all customers ("tenants"). To support scalability, the application can be installed on multiple machines (called horizontal scaling). In some cases, a second version of the application is set up to offer a select group of customers access to pre-release versions of the applications (e.g., a beta version) for testing purposes. This is contrasted with traditional software, where multiple physical copies of the software — each potentially of a different version, with a potentially different configuration, and often customized — are installed across various customer sites.[21]

Although an exception rather than the norm, some SaaS providers do not use multitenancy or use other mechanisms—such as virtualization—to cost-effectively manage a large number of customers in place of multitenancy.[22] Whether multitenancy is a necessary component for software as a service is debatable.[23]

Horizontal SaaS and vertical SaaS are different models of cloud computing services.[24] Horizontal SaaS targets a broad variety of customers, generally without regard to their industry. Some popular examples of horizontal SaaS vendors are Salesforce and HubSpot. Vertical SaaS, on the other hand, refers to a niche market targeting a narrower variety of customers to meet their specific requirements.[25]

Although not all software-as-a-service applications share all the following traits, the characteristics below are common among many of them:

Several important changes to the software market and technology landscape have facilitated the acceptance and growth of SaaS:

• The growing use of web-based user interfaces by applications, along with the proliferation of associated practices (e.g., web design), continuously decreased the need for traditional client-server applications. Consequently, traditional software vendor's investment in software based on fat clients has become a disadvantage (mandating ongoing support), opening the door for new software vendors' offering a user experience perceived as more "modern".
• The standardization of web page technologies (HTML, JavaScript, CSS), the increasing popularity of web development as a practice, and the introduction and ubiquity of web application frameworks like Ruby on Rails or Laravel (PHP) gradually reduced the cost of developing new software services and enabled new providers to challenge traditional vendors.
• The increasing penetration of broadband Internet access enabled remote centrally hosted applications to offer speed comparable to on-premises software.
• The standardization of the HTTPS protocol as part of the web stack provided universally available lightweight security that is sufficient for most everyday applications.
• The introduction and wide acceptance of lightweight integration protocols such as Representational State Transfer (REST) and SOAP enabled affordable integration between SaaS applications (residing in the cloud) with internal applications over wide area networks and with other SaaS applications.

Some limitations slow down the acceptance of SaaS and prohibit it from being used in some cases:

• Because data is stored on the vendor's servers, data security becomes an issue.[32][33]
• SaaS applications are hosted in the cloud, far away from the application users. This introduces latency into the environment; for example, the SaaS model is not suitable for applications that demand response times in milliseconds (OLTP).
• Multi-tenant architectures, which drive cost efficiency for service providers, limit customization of applications for large clients, inhibiting such applications from being used in scenarios (applicable mostly to large enterprises) for which such customization is necessary.
• Some business applications require access to or integration with customers' current data. When such data are large in volume or sensitive (e.g. end-user's personal information), integrating them with remotely hosted software can be costly or risky, or can conflict with data governance regulations.
• Constitutional search/seizure warrant laws do not protect all forms of SaaS dynamically stored data. The result is that a link is added to the chain of security where access to the data, and, by extension, misuse of these data, are limited only by the assumed honesty of third parties or government agencies able to access the data on their recognizance.[34][35][36][37]
• Switching SaaS vendors may involve the slow and difficult task of transferring very large data files over the Internet.
• Organizations that adopt SaaS may find they are forced into adopting new versions, which might result in unforeseen training costs, an increase in the probability that a user might make an error or instability from bugs in the newer software.
• Should the vendor of the software go out of business or suddenly EOL the software, the user may lose access to their software unexpectedly, which could destabilize their organization's current and future projects, as well as leave the user with older data they can no longer access or modify.
• Relying on an Internet connection means that data is transferred to and from a SaaS firm at Internet speeds, rather than the potentially higher speeds of a firm's internal network.[38]
• The Ability of the SaaS hosting company to guarantee the uptime level agreed in the SLA (Service Level Agreement)
• The reliance on SaaS applications and services can lead to SaaS sprawl within enterprises. These disparate applications and services can become challenging to maintain technically and administratively, leading to the proliferation of shadow IT. [39] [40]

The standard model also has limitations:

• Compatibility with hardware, other software, and operating systems.[41]
• Licensing and compliance problems (unauthorized copies of the software program putting the organization at risk of fines or litigation).
• Maintenance, support, and patch revision processes.

Software as a service data escrow is the process of keeping a copy of critical software-as-a-service application data with an independent third party. Similar to source code escrow, where critical software source code is stored with an independent third party, SaaS data escrow applies the same logic to the data within a SaaS application. It allows companies to protect and ensure all the data that resides within SaaS applications, protecting against data loss.[43]

There are many and varied reasons for considering SaaS data escrow including concerns about vendor bankruptcy,[44][45] unplanned service outages, and potential data loss or corruption. Many businesses either ensure that they are complying with their data governance standards or try to enhance their reporting and business analytics against their SaaS data.

One notable criticism of SaaS comes from Richard Stallman of the Free Software Foundation, who refers to it as Service as a Software Substitute (SaaSS).[46] He considers the use of SaaSS to be a violation of the principles of free software.[47] According to Stallman:

With SaaSS, the users do not have even the executable file that does their computing: it is on someone else's server, where the users can't see or touch it. Thus it is impossible for them to ascertain what it really does, and impossible to change it.

• Application security
• Application service provider
• Cloud-based integration
• List of 'as a service' service types
• Servicizing
• Subscription computing