SDWAN Solution- Does it really bring cost saving!

Proponents of #SDWAN are coming up with news articles on marvel-of-technology nearly every day, predicting explosive growth, promising new world of branch-in-a-box and loads of never before provider centric and customer centric benefits. Having reviewed many of these news articles, blogposts and analyst reports from subject matter experts in the last 12 months, it seems most of them are presenting a helicopter view of technology trend, market landscape and how SDWAN solution ushers a new era of network-as-a-service. However SDWAN solution delivered as end-to-end service to customers merits a closer look as the perspective of SDWAN technology vendors and SDWAN service providers could differ, at times considerably. Hence this article explores the ground reality of delivering SDWAN solution and evaluates the potential of cost saving, in a global landscape.

Let’s start by considering the customer use case that is most suited to adopt SDWAN solution. Such a customer is likely to have an enterprise WAN solution with MPLS and/or Internet paired in primary-backup configuration and Cisco CPE taking charge of failover switching. This customer is interested in SDWAN solution, provided it leads to cost saving or at best a minimal increment in MRC. This of course must be in addition to regular year-on-year renewal discounts for existing network services, which the customers consider as their fundamental right.

First, who can provide SDWAN solution with Cisco CPE. The most likely option is Cisco-Viptela solution that was launched recently. The extent of service integration is not clear and it is expected that Cisco will progressively integrate Viptela solution with their ISR, ASR and ENCS platforms. The other option is from Tata Communications, that has a patent pending solution to leverage the existing Cisco routers. However for a SDWAN service provider, there is lot more to it in order to put together a SDWAN solution, technically and commercially.

Commercially, SDWAN overlay will lead to MRC increment for every site, and that may need to be off-set by rearchitecting the underlay. This could be achieved by MPLS bandwidth downgrade and Internet bandwidth upgrade. However, it should be noted that in most markets the price differential between MPLS and Internet is narrow and tapering progressively, as both the services have similar network construct and outflows. Alternately if the customer is planning for bandwidth upgrade to cater to growing application workload, SDWAN solution could avoid or minimize such upgrades leading to cost avoidance.

The commercial jugglery to construct a cost saving proposition is easier in the home country of the provider where nearly all the network components of the end-to-end SDWAN solution is owned and operated by the SDWAN provider – MPLS service, Internet service, Local Access, SDWAN overlay with end-point appliance, service gateways and centralized controller. The concept of home country could however be different in other geographies like Europe where borderless business economy has led to the cluster of countries become one mega-country. That’s the likely reason growth of SDWAN service is more likely in the home market, in the initial period of service launch. Global expansion for sure can be and is pursued in parallel, starting with well-known developed markets followed by select emerging markets.

Technically, SDWAN solution gets challenging outside the home country where the SDWAN provider either has licensed entity with fully owned PoP setup or a partnership arrangement with a local licensed provider. The partnership arrangement for high potential countries is likely to have MPLS PoP and IP PoP setup and for others it could be either MPLS PoP or IP PoP setup depending on product roadmap. In countries where only partnership MPLS POP is available, the Internet services are usually procured from local ISP or global providers like Expereo, Brodynt or Globalinter.net. In countries where only partnership IP PoP is available, the MPLS service is procured through NNI with partnered network service provider. These combinations are very well suited for Hybrid VPN solution, however they are not suited for SDWAN solution or requires highly customized network engineering for the  solution to work. This restricts SDWAN solution to be offered in countries where MPLS and IP PoPs are available. Hopefully propelled by the need to expand the addressable market, and advanced network engineering, the technical challenges can be overcome. Several providers are reportedly working on it aggressively and expected to make announcements later this year.

Technically, there is one more challenge to be considered. In high potential target countries outside the home market, developed or emerging, the SDWAN provider needs to extend the SDWAN network. This is done by setting up one or more SDWAN service gateway. I’m not elaborating the purpose and benefits of SDWAN service gateways and presume readers would be aware of it. The SDWAN service gateway setup has it technical, commercial and even contractual requirements. The SDWAN appliance at customer premise in a given country will communicate with the SDWAN service gateway that will in turn have upstream connectivity over MPLS and IP network to the central controller, usually hosted in the home country. That’s more so for multi-tenanted SDWAN platforms. For uni-tenanted SDWAN platforms, every customer SDWAN network is a self-contained solution. It seems multi-tenanted SDWAN platforms have several advantages over uni-tenanted SDWAN platforms. Finally, the SDWAN service gateway needs ongoing support with spare stocking, 24/7 fault-fix and RMA for the hardware. The SDWAN provider is responsible to make the necessary arrangements for it.  

There is another option for cost saving, one that is generally advocated by SDWAN technology vendors. Replace private carrier service like MPLS with Business Internet delivered with metro fiber network and implement SDWAN in conjunction with Consumer Internet or broadband Internet delivered with 4G/LTE. It is likely to bring some cost saving. However is the customer agreeable to replace the trusted MPLS links with Internet given all the disadvantages of a best-effort public network. Cost optimization mandates might compel some customers to consider it for lower-workload branch offices. Further such replacement is more practicable operationally in the home country, provided the replacement is provided by the SDWAN service provider. In emerging market countries getting distinct flavors of Business Internet and Consumer Internet could prove challenging.

Let’s look at another customer use case, where the customer has MPLS and/or Internet, and is agreeable to replace the fleet of Cisco routers with SDWAN appliance with well-orchestrated migration plan. In the home country it really works well. SDWAN provider will have its field operations team and alliance with a logistic partner to stock, deliver and install the SDWAN appliance at customer premise and provide on-site replacement within 4-8 hours if it fails. However outside the home country, additional arrangements are necessary. One or more global logistic partner is needed for selective global coverage. Import regulations of many countries like Russia, Brazil and Indonesia to name a few are quite challenging to navigate. Local procurement, if somehow feasible, will be more costly. Further the global logistic partner needs to have or stitch up alliance with a local support partner for target countries. This is quite unlike the experience with Cisco SmartNet support and its global availability. The SDWAN technology vendors have not made efforts till date to establish global logistics and support partners. Hence each SDWAN service provider is compelled to appoint its own logistics and support partner and manage it operationally. Not always does it work predictably and challenges often are country specific. Here again with experience SDWAN providers will become competent to manage delivery and assurance of SDWAN solution realiably with better foresign on possible challenges.

With so much of efforts for end-to-end delivery of #SDWAN solution on a global landscape, is it really practicable to offer double digit cost saving. I welcome views from subject matter experts in SDWAN to share more experiential insight and case studies. 

(The views expressed by the author in this article are his own and not necessarily relates to the organization the author is attached to)