Generally, organisations have chosen one of two models for creating a Wide Area Network.  The safest model has always been to buy circuits from a Tier 1 telco, in managed or wires-only configuration, such as MPLS from BT – relatively secure, scaled according to site bandwidth and a reliable, SLA-backed, service but with a considerable price tag.  True resilience was always achieved by deploying alternate circuits from a secondary Telco, ideally using diverse routing to minimise single points of failure (SPOFs) but, in effect, doubling the costs without an optimum use of bandwidth.  Alternatively, DSL or cable services have the benefit of lower cost and potentially faster speeds but also present reliability and resilience issues.

 

The balancing act between price and performance has been a significant challenge for most organisations but cloud is also re-writing the network rule-book as people regularly consume business technology outside the traditional WAN and the security perimeter has now been re-defined.  Now is a good time to review WAN design and seek better design models as bandwidth costs have fallen and cloud consumption has increased.

 

Hybrid WAN allows organisations to reconsider network design, blending reliability from reduced cost Ethernet circuits with fast bandwidth on DSL, fibre and even 4G/5G services.  Tailored design of services available to match each organisation’s needs can now produce the same levels of performance as traditional (and expensive) corporate WANs and can deliver the best of both worlds and eliminating SPOFs at the same time.

 

By using hybrid WAN technologies, organisations can create policies to route traffic down multiple network connections according to security requirements and/or load. The policies can be driven by application type, performance, cost or any other criteria the organisations might deem important.  Rapid deployment can also be achieved and integrated by use of 4G/5G wireless services.   A dynamic hybrid WAN solution can continuously monitor the network and send traffic down the best-performing link whilst still being governed by policies.

 


PDG Diagram incl SPOFs and Locations 

 

This brings a level of agility to the WAN that has never been seen before and is a critical step in the network’s evolution to become ‘cloud capable’.  In a traditional hub-and-spoke network, cloud traffic would come from the Internet into the DataCentre and then up the spoke to a branch office.  Each of these hops adds unnecessary latency to the flow of traffic, meaning that software-as-a- service (SaaS) applications don’t perform as well as they could.  Now that the cloud is being used for applications that require high- performance networks—such as voice, video, VDI and data replication—it’s become crucial that the network is designed ‘cloud first’ but there are other benefits too:

 

Reduced cost: Hybrid WAN can, in some cases, eliminate MPLS.  Conservative enterprises that want to migrate away in stages can at least reduce the amount of their budget currently spent on MPLS.

 

Improved application performance: In Hybrid WAN, all the bandwidth from each connection is active. Even though Internet traffic doesn’t have the same ‘guarantees’ as MPLS, multiple active connections protect against outages and ensure optimal performance.  With multiple active routes, organisations are likely to find that Hybrid WAN provides similar availability metrics as MPLS.

 

Improved security: Contrary to popular belief, an MPLS network is not 100% secure. Because traffic is unencrypted, any network breach means malware can spread to any point on the network quickly. When Internet connections are used, the traffic is typically encrypted from the source to the destination, potentially offering better data protection, dependent on control measures in place.

 

When considering Hybrid WAN design, it’s important to consider that the design is much more than a conglomeration of connectivity; the real advantage lies in the capability to design and create policies based on application types, application performance, security and cost and must include the ability to dynamically adjust as the network conditions change, so monitoring is an essential part of the design. 

 

Wireless networks are changing the way people communicate and access information. The mobile access network, enabling connections anytime and anywhere, is transforming the technology industry. 5G is the next evolutionary step; providing flexible, reliable, and secure wireless networks to connect people with all applications, services, and things heralding the ‘everything on mobile’ age. Key 5G characteristics include: 

 

Peak data rate per user: 10 Gbps - 4 times of the capability of current 4G network

 

End-to-end latency: 1 ms - 1/50 of best latency of current 4G network

 

Number of connections: 1 million per square kilometre - 100 times the current 4G network

 

The potential for 5G is immense but 4G is here right now and can already facilitate rapid WAN deployment and/or temporary extensions at usable LAN speed, dependent on application and requirement. In some cases, this is actually preferable to DSL but also can offer true resilience when used in conjunction with another medium.

 

Cloud computing has revolutionised technology over the last decade and it will continue to significantly change the way organisations operate and how people work. Hybrid WAN has the potential to underpin cloud whilst providing cost and performance efficiencies. Design has always been important in technology but never more so than now.