Infrastructure as a Service (IaaS) explained
Software-Defined Networking (SDN); Network Functions Virtualization (NFV); Infrastructure as a Service (IaaS); Platform as a Service (PaaS); Software as a Service (SaaS). If you have been in technology for any length of time, you understand that you will be immersed in mnemonic soup. However, these are new terms and they represent a significant paradigm shift in the world of networking. This shift is so disruptive that it has shaken the core of every service provider and Network Equipment Manufacturer (NEM), from Tier 1s to startups. Of course, disruptive technology and paradigm shifts signal vast opportunity, and also threaten to unseat those on top. This column will briefly examine this paradigm shift and a key component of it relevant to the ATCA world – Infrastructure as a Service (IaaS).
What is the paradigm shift?
The short answer is that the paradigm shift involves an evolution of network infrastructure from a variety of special-purpose platforms with specific functions to a common virtualized platform architecture consisting of standard high-volume servers, high-volume storage, and high-volume Ethernet switches. Industry calls this SDN.
The benefits are:
- Lower capital expense for equipment since everything is the same and economies of scale can be achieved. Large numbers of standard high-volume servers can be deployed as the network infrastructure.
- Faster rollout of new features. Independent software vendors can quickly write, test, and deploy network applications.
- More innovation as independent software developers can write for a common, virtualized environment.
One of the key tenants of SDN is the separation of the control plane (network protocol layers that make decisions about where packets go) from the data plane, which performs the nuts-and-bolts forwarding of the packets. Since SDN decouples the control and data planes, there must be an interface that allows communication from the control plane to the data plane. One specification that has arisen that addresses this is called OpenFlow.
In short, the Tier 1 service providers are taking a page from datacenters and plan to begin transforming their networks into one big datacenter whose money making application is the transport and processing of network traffic.
One of the complementary capabilities of the datacenter is virtualization (Figure 1). Servers running virtualized environments provide the ability for network applications to be developed, tested, and run anywhere that virtualized environments can be hosted. Virtualization also provides for a level of security since virtualization environments are separated and cannot affect each other. The networking world is calling this NFV.
Relevant standards and organizations
The Open Networking Foundation (ONF) is a trade association that was founded to improve networking through SDN. The founding principal was based on the fact that cloud computing further blurs the line between computers and networks.
The European Telecommunications Standards Institute (ETSI) is also playing a key role. At the October 2012 SDN and OpenFlow World Congress, ETSI published an introductory white paper on NFV and its positioning relative to SDN, OpenFlow, and cloud computing trends. It is an excellent read for those interested.
How ‘X’aaS fits into the paradigm shift
Cloud computing has three general business models:
- Infrastructure as a Service. This business model sells processing power, storage, and computing power – basically, the server, storage, networks, and Operating Systems (OS) baseline. It is the pallet upon which all other services are built.
- Platform as a Service. In this model, the supplier provides a set of tools and services designed to make coding and deploying applications fast and easy. This is the next layer up in capabilities from IaaS.
- Software as a Service. This service is where the provider develops applications for end users and delivers them over the web. The application software and data are hosted in the cloud and is accessed by users with some kind of thin client or web browser.
IaaS and ATCA
ATCA certainly seems well positioned to be the Commercial Off-The-Shelf (COTS) platform of choice for these deployments given its tailored architecture toward telecom, commercially available compute blades, and switching architecture. Some naysayers believe that SDN means less reliability due to what they believe will be enterprise-level servers proliferating throughout the network infrastructure. However, the NEBS compliance and platform management features within ATCA, in conjunction with compute blades, I/O blades, storage, and switch fabric, provide a reliable, interoperable foundation for SDN and NFV.
Prospects of a paradigm shift
These are extremely exciting times for the communications industry. The prospect of a new age where software innovators can quickly create and deploy solutions for consumer problems is fascinating. Standards bodies are backing the movement, specifications are being produced, and labs around the world are advancing the technology. The SDN/NFV wave promises to be the single biggest inflection point since the Internet became widely available to the public.