By now, you have probably stumbled upon the term Software-Defined Wide Area Network (or SD-WAN). But SD-WAN is not just another buzzword, and it has quickly gained traction among CIOs. The reason being, SD-WAN enables CIOs to confidently consider hybrid WAN deployments that make the best of the ubiquitous and inexpensive nature of the Internet and the privacy and reliability of MPLS links.
MPLS VPNs are the dominant type of WAN service used by enterprises that need to connect their distributed enterprises globally. MPLS connections are private networks that never touch the Internet, and offer a high level of reliability and performance, as they enable traffic prioritisation using the class of service (CoS) feature.
However, a MPLS-based WAN architecture is static in nature, as the network is configured using dedicated proprietary hardware, i.e. routers and switches, at all the sites that need to be on the WAN. This static nature of the architecture adds to lengthy circuit provisioning times and complexity in configuring and managing connections across thousands of enterprise sites worldwide. MPLS circuits are also expensive, as the cost involves a proprietary router at each site, along with the cost for access circuits, bandwidth, and the associated CoS fee.
Furthermore, an MPLS WAN architecture assumes that all business-critical applications reside in enterprise datacentres, which branches connect to a private WAN. While MPLS networks are typically architected in a fully meshed model, branch network traffic is still routed via a headquarter or hub location to ensure data security and compliance, which adds delays. The increased enterprise use of cloud and mobility applications means that this model is not fit for purpose anymore, as the majority of traffic is no longer exchanged within the private datacentres of the enterprise. Cloud applications are often accessed over Internet links; and the number of mobile users connecting over the Internet to enterprise applications has grown significantly. Hence, the “hairpin” model of traffic being routed via the enterprise datacentre no longer applies.
An SD-WAN architecture automates the process of network selection so that both private WAN and public Internet can be utilised efficiently. While MPLS VPNs are provider-managed – meaning the CIO and network administrators have little control over the management of the network – SD-WAN puts the control back in their hands, while enabling the enterprise to use a combination of private and public networks. This results in a better total cost of ownership (TCO). Lower costs and the ability to ensure optimal application performance through greater control of branch sites remotely has attracted significant interest from enterprises, especially from verticals with distributed branch locations, such as retail, banking, financial services, manufacturing and logistics.
In a recent Frost & Sullivan end-user survey, 18% of the respondents said they are already using SD-WAN, and 25% of them indicated they will deploy SD-WAN in the next 12-24 months. Among the respondents planning to deploy SD-WAN in the next 12-24 months, a whopping 57% said they would replace 50% of their MPLS connected sites with SD-WAN.
Figures like this highlight that enterprises already have aggressive deployment plans for a solution that is less than two years old. In tandem, SD-WAN vendors and network service providers are gearing up – forging strategic partnerships and building internal expertise to deploy and manage SD-WAN solutions – to fulfil the growing demand for SD-WAN.
How do you see SD-WAN transforming IT in global organisations? Let us know in the comments below.