Continued from Page 1
If you are global (or even regional) in nature, no single service provider will be able to provide the lowest cost BB to all your locations. In fact, for some locations it's likely that service providers will not be able to provide BB at all. Managing the number of service providers is a cost versus complexity trade-off. Unless having one service provider to manage (aka one throat to choke) is your primary requirement, you should plan to engage with several as a minimum. Aim to create a portfolio of cost-effective, trusted providers.
Additionally, you should expect to have some "exceptions" where local service providers fill gaps based on cost, coverage, or speed grounds. If a single provider is really what you want, note that this can (and will) come at a significant cost premium with a total cost differential that few enterprises would want to leave on the table. Interestingly, DIA is more suited to single provider provision as traditional carriers are better able to rely on their own infrastructure to provide "on-net" service and see greater potential to justify their "value add" given the higher price of DIA.
Before you cry foul, we are aware of the "aggregators" that some see as the panacea to their Internet transport management woes (read our previous No Jitter post, "Will Broadband Aggregation Reach Critical Mass This Time?"). Unfortunately, for BB (and DIA for that matter), the challenges relating to cost, coverage, and speed still apply with aggregators -- as do some new challenges around thinner account teams and, as yet, less flexibility in establishing contracts.
Much to Think Through
You must consider the challenges of applying enterprises' traditional heavy-duty master service agreements and schedules with some service providers that are used to simply using service order forms. Plus, the underlying BB providers (used by aggregators or via carriers' wholesale arrangements) are not used to or willing to have enterprise terms and conditions cascaded down to them, even if the precedent, time, and appetite existed for your chosen provider(s) to attempt this. Extensively using BB involves recognizing and accepting more contractual/commercial risk -- for example, don't waste time trying to cascade terms down to all underlying providers, but rather work with your providers on the "art of the possible" obligations that are most likely to affect service.
Variable performance parameters such as delay and jitter are also potential problems with any service that is largely delivered via the Internet. With contended access, very location dependent offerings, and asymmetric upload and download speeds, BB performance is an issue to consider. Most affected will be delay-sensitive applications such as voice and video. The counter to this is the general user experience of Skype and other such services over own home broadband. Service can be affected but can be "good enough" and user expectation (less concern about interruptions) has opened the way to putting such services over BB. Again, the impact of SD-WAN for improved performance is relevant here. That said, there will always be greater risk of poor performance, particularly in times of Internet congestion.
For higher quality services, the case for using BB becomes more fragile. BB uplink speeds are constrained to a few megabits per second in many areas. In China, for example, BB speeds are limited to 1 Mbps upload. The experience for some users may not be adequate compared to those blessed with the highest-speed BB. While applications can be prioritized, they can still have an impact on other services.
Bandwidth (i.e., circuit speed) is another issue to think through. The available speeds for BB are highly variable globally, regionally, and in the same country. As mentioned, some countries have severe limitations that immediately drive enterprises to use DIA for Internet transport. In other cases, even in areas with 50 to 100 Mbps BB, and more so where available speeds are 5 Mbps, 10 Mbps, 20 Mbps, etc., the available maximum uplink speeds are the primary constraint, the BB circuits are simply not up to the total bandwidth demand. Here, the DIA option may need consideration. In procuring your BB, you need to be flexible in what you ask for and what you are prepared to use. Asking for a standard speed globally won't work. You need to understand what your minimum uplink speeds are and as a first step focus on that; thereafter you can also explore what higher speed services are available. This requires more complex analysis and assessment of where BB might work for you.
Ultimately, to maximize the use of BB, an effective SD-WAN solution is a pre-requisite. Many enterprises are not there yet. There are pitfalls in SD-WAN solution selection, harmonization of implementation with your transport acquisition, and in considering how best to execute on your network transformation. You may also have existing contract spend or circuit obligations that will potentially affect transition, although the ever-growing bandwidth demand may make BB deployment possible on "cost avoidance" grounds -- for example, to avoid perpetuating or increasing spend on MPLS.
Overall, the extra planning and analysis is likely to be worth it, aswill the change in practices to accommodate multiple and different suppliers. The variability in BB availability, performance, speed, SLAs, and contracting should not prevent you considering extensive use of BB as part of your network transport portfolio. With BB being an order of magnitude lower cost than MPLS or DIA circuits (often lower-bandwidth circuits) the cost control opportunity can be very significant.
Mark Sheard is managing director TC2(UK). Based in London but working globally, Mark is part of TechCaliber Consulting, LLC, a global IT and telecom consultancy that advises the world's largest companies on transformational strategies for reducing their costs for telecom and IT products and services. Mark can be reached at [email protected].
