Redundancy for Corporate Broadband WHITE PAPER December, 2012
Introduction Over the past twenty years, broadband service providers have been using different mediums to deliver their services. From wireless to wired (fiber, coax, etc.) the pros and cons for each category have been discussed in numerous papers and articles extensively. But there is not enough discussion about how wireless and wired broadband delivered together can increase the reliability for the end user. In this paper, we discuss why aggregating your wired broadband with Skyriver wireless can increase your broadband reliability and uptime. Telecommunication Network Deployment Telecommunication networks have evolved dramatically in the past couple of decades. From twisted pair connections like T1s, to higher speed connections like DS3 and fiber optics, customers have many wired choices for their broadband needs. However, all of them have one thing in common, they are laid in trenches. Common Threat for Wired Telecommunication Networks As discussed in the previous section, the close proximity of trenches and cable/fiber runs for legacy wired telecommunication operators makes them all vulnerable to the same physical and natural accidents and disasters. A flash flood in one part of the city can wash away part of a street or a highway, or an underground gas pipe explosion can destroy all trenches and cable in the vicinity causing network outages for all operators with cables/fibers in that area. Reconstruction after events will be lengthy and cause long outages for the affected neighbourhoods even miles away if their feeds go through destroyed trenches. Therefore, the concept of providing redundancy for critical enterprise networks by adding multiple wired providers is not logical. In case of any natural or man made accidents, all operators networks in the area will be affected since all of them use the same foot print. Most wired operators lay cables or fibers in specific paths under roads across the country. Because of the difficulties in construction and coordination with cities and local authorities, most operators either share the same trenches, or lay their cable in their own trenches which are next to the other trenches. These issues will cause long lead times for wired operators to provide services in areas they do not have presence. 2
Why Fixed Wireless Networks Are Different In a fixed wireless network, a base station can service hundreds of dissimilar subscribers in terms of bandwidth and services, as shown in Figure 1. Since the medium is wireless, any physical damages or changes on the ground between the base station and customer premises will not affect the network. While major destruction destroys trenches and roads, the wireless connections remain solid. Sectoral Antenna Figure 1: Point-to-Multipoint Fixed Wireless 3
Skyriver Maintains Upstream And Access Diversity Like major telecommunication companies, Skyriver utilizes a multiple upstream approach. In each market, Skyriver s core network connects to at least two major upstream providers, using a distributed BGP (Border Gateway Protocol) design, to provide complete redundancy for Internet access. This eliminates unexpected outages due to failure from one of the upstream providers. Businesses Utilize Path Diversity For Full Redundancy Businesses with mission critical operations looking for full network redundancy must combine their wired connection with a fixed wireless connection to create a truly redundant network with network diversity and path diversity. A Skyriver fixed wireless link will bypass all surrounding street level hazards and guarantee connectivity to the network should a major outage occur for wired operators due to a natural or man made disaster at the street level. While wired operators go through a lengthy reconstruction of their network, which can take days, if not weeks, a business utilizing Skyriver will be up and running via the Skyriver link. Carrier diversity is also important because if one provider goes down for any reason, you are still connected by another provider, thereby adding maximum reliability. Achieving Full Redundancy With Failover and Load Balancing Full redundancy is made possible by using a router with multiple WAN ports that provides for failover and traffic load balancing. There are a number of brands to choose from and many factors to consider when selecting a router. You can find units with 2, 3, 4, and up to 30 WAN interfaces. Each WAN port can connect to an ISP. When selecting a unit, keep in mind any growth that your company may have over the next few years. Note that the throughput is the maximum inbound and outbound speed the unit can support simultaneously. For example, a throughput of 40 Mbps means that the unit can do 20 Mbps down and 20 Mbps up at the same time. Get good clarification from the vendor on this, as well as confirm that the number of NAT sessions the router supports simultaneously meets your needs without slowing down performance. 4
Failover Multi-WAN routers have the ability to re-route your inbound and outbound ISP traffic in the case of a failure. You ll want to set-up your failover procedures when you set-up your multi-wan router. By setting it up in advance, you can avoid down time since failover procedures will automatically re-route traffic to another healthy connection. Take care that the router you choose provides for adequate balancing policies for the type of traffic you have. IPs issued by carriers are all carrier specific. So the loss of a primary circuit will affect IP schemes including the availability of self-hosted public facing applications such as email, VPN, collaboration portals, content servers and VoIP. True failover or backup capabilities need to account for DNS (Domain Name System) in the network design. In the event of ISP unavailability, the DNS entries for those external IP addresses will need to redirect traffic to the IP addresses of the backup ISP. This can be done manually or through the use of automatic methods provided as a service by Dynamic DNS hosting providers or features built into the load balancing router. Load Balancing Load balancing is a technology that distributes network traffic over multiple active WAN links. Load balancers are typically more expensive since they do both outbound and inbound load balancing and they inherently provide failover. It is important that your unit be able to force (not balance) specific traffic, as in the case of mail servers, VPNs or https/443 traffic as these require non-fluctuating IPs. Additionally, other institutions, such as banks will do secure traffic checking and will block an IP address that changes in the middle of the transaction. Therefore, you need to be able to force the device to not load balance for these types of websites. You want to have granular control over balancing traffic based on criteria such as source, destination, port and time of day. Inbound load balancing is generally performed via dynamic DNS that can either be built into the system or provided by an external service or system. Dynamic DNS built within the device is generally better from a cost savings and overall control point of view. Some dual-wan routers have a built-in DNS server to provide inbound load balancing. 5
Conclusion For businesses looking for maximum reliability and uptime, aggregating reliable fixed wireless and wired broadband services is ideal. Redundant broadband services, utilizing a separate and distinct fixed wireless provider plus a wired provider will ensure true path and carrier diversity, thereby increasing dependability and reliability for mission critical operations. 2012 Skyriver. All rights reserved. 6