This site, like many others, has focused on the need to evolve the WAN. Much of the value proposition for the shift to a hybrid WAN is to improve application performance, which is often poor over WAN. However, the challenges with a legacy WAN go beyond application performance. Operational issues are as big or a bigger problem than the application issues, they’re just not as noticeable because they only impact the network operations team. Here are what I consider to be the most significant operational challenges with legacy WANs:
- Long lead times in upgrading bandwidth. The length of time it takes to upgrade T1 MPLS circuits can be 30 to 90 days at a minimum. I’ve actually seen an upgrade of a T1 take as long as six months when the local telco had to come out and upgrade the facilities in the building. When higher speed links are used, like DS3s or OC3s, the process can take even longer because of a lack of fiber. Internet connections can generally be upgraded much faster and if some kind of WAN aggregation device is used, a 4G connection could be added for an instant bandwidth upgrade.
- Inefficient use of bandwidth. The current architecture used to build a WAN with multiple connections is to have one connection active and the secondary connection passive and only becomes active with a failure of the primary link. This is done to prevent routing loops and broadcast storms. Obviously having to buy two connections and having one always in a passive state is a huge waste of money. Also, if a “hub and spoke” architecture is used where Internet traffic is sent through a single choke point, the same traffic traverses the WAN twice. Because of this, network managers are constantly fiddling with network settings, management tools and optimization platforms to try use the network more efficiently.
- Managing branch infrastructure. This is one of the biggest challenges for network managers. Managing remote infrastructure is tough. It’s even tougher when there are multiple appliances used to deliver the necessary functionality. This might include routers, security devices, WAN optimization appliances and other network functions. Often making changes to the WAN requires an IT person to visit the branch. In a large, distributed enterprise, it can literally take years to get through a WAN upgrade because each location requires on site maintenance. The rise of network functions virtualization (NFV) as a cornerstone of the hybrid WAN should greatly help here. No more truck rolls, no more asking the local administrator to find the router and power it on or off and no more fiddling the configurations to ensure one services works with the other.
- Lack of application visibility and control. I’ve always said that you can’t manage what you can’t see, and it seems network managers are seeing less and less every year. The rise of cloud and virtual platforms means that there are more “blind spots” on the network than ever before. Also, legacy WAN architectures and routing protocols do not provide any level of application control. This makes mapping business policies to network policies difficult, if not impossible to do. A software-defined WAN coupled with virtual WAN optimization services enables the application needs to automatically make the necessary WAN changes to provide better control. Also, many of the new network management tools look at traffic at a flow level so virtual and cloud traffic can be mapped into the physical network.
Running a WAN has never been an easy task and it seems all the advancements in technology have made the process even more complex. A next generation, hybrid WAN that utilizes SDNs and is architected for cloud traffic and greatly reduce the operational overhead associated with WAN management.
This post is part of an ongoing series examining the issues facing enterprises seeking to implement a Software-Defined WAN (SD-WAN) solution, as addressed in the Open Networking User Group white paper, “ONUG Software-Defined WAN Use Case”.
Image credit: felixtriller (flickr) / CC-BY