Infrastructure-as-a-service (IaaS), the fastest growing segment of the rapidly expanding cloud market, is going mainstream into the enterprise. Concurrently service providers, who were early adopters, are moving upstream to more higher-value services
IaaS is a simple concept, but with complexities that must be dealt with to successfully implement. The basic idea is to deliver data processing services quickly on a converged infrastructure that pools resources and automates repetitive manual processes to speed time to market. It minimizes redundancies, though initially it can be disruptive to how infrastructure is managed because it aggregates compute, network, and storage resources, moving the autonomy enjoyed by domain administrators (e.g. storage admins) to a central function.
IaaS first found acceptance with service providers looking to leverage their data center investments to provide cloud services for new revenue streams. As technology evolved, and cloud economics were better understood, enterprises began to apply the IaaS model within the data center. Several approaches have emerged, including single-SKU solutions that give service providers and enterprises the ability to roll-in prepackaged infrastructure to rapidly deploy services whether to end-user customers or internal business groups.
From Time-Sharing to the Cloud
IaaS is not a new concept. In the age of mainframes, starting in the 1960s and continuing into the 1980s, applications ran on one time-shared central computer system, which everybody accessed via terminals. Within the enterprise, IT had to offer services to ensure effective management of the shared compute resources.
Externally, many service providers offered basic accounting applications to enterprises. A great example of a time-shared application is ADP payroll services which is still used by many organizations today, including EMC. These service providers, including ADP, also offered excess capacity for hosting or developing applications. In the 1990s, the rise of personal computing allowed applications to run on individual compute resources, and the notion of renting IT services moved to the background. Then, the Internet simplified data transport and access and commodity-based compute, network, and storage introduced different economics which has brought back the concept of IaaS.
Initially, in this rebirth, managed-service or cloud providers such as Amazon, CSC, and Terremark emerged as early IaaS adopters. They provide the hardware infrastructure and associated software to end-user customers, as well as the management and maintenance on a per-use basis. Converged infrastructure, like the VCE Vblock system, helps lower hardware and support costs, making their IaaS offerings even more attractive to their customers. This approach seems to still be holding true, as the time-to-market and costs of providing IaaS are still out of reach for many small-to-mid size businesses (SMBs). But, there is a new trend emerging.
From the Public Cloud to the Enterprise
IaaS is moving from public cloud to larger enterprises. This change comes as no surprise as CIOs, now more comfortable with cloud technologies and economics, look to bring the process efficiencies, business agility, and cost savings in-house. Or, alternatively, have the option to provide a combination of on-premise and outsourced computing—with the hybrid cloud.
Industry-analysts The Gartner Group forecast IaaS to be the fastest growing segment in the cloud market. Revenues are projected to grow from $5.6B ($USD) in 2011 to $24.4B in 2016. Technology innovations like virtualization and converged infrastructure management software are giving large enterprises the tools necessary to provide IaaS internally and often at a lower cost than public cloud service providers.
IaaS in the Real World
To illustrate the possibilities, let’s look at some real-world examples of enterprise and service provider (for hybrid cloud) IaaS deployments, including the challenges each organization faced:
- Molina Healthcare: Difficult economic conditions have generated skyrocketing demand for US government health programs. As a result, this Fortune-500 healthcare provider has seen huge increases in claims processing loads and rapid IT infrastructure growth. Their converged infrastructure solution approach to IaaS gave them the ability to launch their private cloud in just a few days.
- Holy Name Medical Center: In this era of flat budgets and high expectations, this regional medical facility in northern New Jersey (US) was running out of physical data center resources, and with their lean staff, was unable to deploy applications fast enough. Now, by utilizing converged infrastructure, their physical constraints are resolved. Additionally, with the new management capabilities that came with their single-SKU solution, they can now deliver infrastructure-as-a-service and roll-out applications in minutes rather than days.
- ATScloud: This US-based managed services provider offers hybrid-cloud solutions to SMBs looking for Fortune 500-caliber IT functionality without the upfront infrastructure and internal staffing investment. Because of their successful business model, ATScloud needed a way to cope with their three-fold (3x) increase in business year over year. The move to a converged infrastructure solution, with its consolidated and fully virtualized environment that let the company accommodate its own growth while providing speedy scalable IaaS to its end-users.
- ReasonNet: As market interest in IaaS grew, service provider ReasonNet in the Netherlands realized that its outdated data center might limit its ability to capture its share of this emerging market. Their solution included pre-integrated and pre-configured infrastructure consisting of compute, network, storage, virtualization, and management layers. The converged infrastructure management software enables ReasonNet to configure, monitor, and troubleshoot the system from a single control point and deliver IaaS.
In each of these examples, the organizations chose a single-SKU Vblock solution with EMC Unified Infrastructure Manager. Alternatively, they could have opted for a tailored approach assembling their own best-of-breed solutions or a vendor reference architecture serving as a blueprint for the IaaS solution. The reference-architecture route would involve assembling orchestration software like VMware vCenter Operations with element managers like EMC Unisphere and EMC Smarts for availability assurance and compliance.
Regardless of the approach, IaaS has moved into the enterprise in private- or hybrid-cloud deployments. With the software-defined data center (SDDC) on the horizon, IaaS can only get more pervasive. The SDDC promotes the concept of moving compute, networking, and storage intelligence into control planes that serve up infrastructure resources on-demand to meet application and service needs. Server virtualization moved compute intelligence into the control plane years ago, making it possible to deliver virtual machines in minutes. With the move to do the same now to networking and storage, presenting infrastructure services can only become less costly and quicker for enterprises and service providers. While enterprises might regain some of the business originally lost to service providers, the market opportunities are still enormous for IaaS. Many service providers tailored their data centers and business models for IaaS, positioning them to offer higher-value services for new revenue streams.