Current Filter: Network>>>>>Feature>

   

Just about every advert you see or hear about internet connectivity focuses on bandwidth and reliability. Everyone wants a faster and more reliable internet connection and ISPs insist that they are working hard to make it happen. The general consensus is that they do just that. ISPs are deploying Fibre to the Cabinet, Ethernet in the first mile (EFM) and upgraded DSL services, and Mobile Service Providers are now deploying higher capacity 4G/LTE networks.

Generally it is much easier for us to get a high speed internet connection, especially in urban and metro locations, at prices we consider fair. However, if you need high and clearly defined reliability as well as bandwidth, then costs begin to rise. ISPs charge more for lower contention on circuits and Service Level Agreements on connection availability, and even more for point to point, dedicated fixed links with SLAs.

The current norm when targets are set for the highest possible bandwidth with the most reliable connectivity is to install a comparatively expensive dedicated fibre connection from the ISPs nearest Point of Presence, to your location, along with a secondary failover circuit, often a DSL or EFM connection. The idea being that if the primary fibre link fails, your internet service will fail-over to the secondary link until the problem is resolved.

The trouble with this scenario is that both the primary and secondary links are normally from the same ISP, meaning that the fail-over is only effective when the fault lies with the fibre link or its associated hardware. If the fault is actually in the ISP's core network, or it is an issue with a key core service like DNS or routing, then both primary and secondary links can be unavailable at the same time. This situation can happen more often than most people realise, and this is a major flaw in this type of fail-over configuration. The bottom line here is that we should expect every internet connection to fail at some point, and plan accordingly. A part of that planning should include a scenario where all connectivity from a single ISP fails at the same time. Understanding the options is vital.

In nearly every other aspect of IT service delivery we are seeing the decoupling of services from their underlying technologies. Server virtualisation and Software as a Service are examples of technologies that decouple the end user service from the underlying architecture to provide both reliability and flexibility that were historically difficult to scale and difficult to recover from when failure occurred.

Reliable and highly available internet access benefits from decoupling internet access as a service, from the underlying connection technologies and their ISPs. It is the combination of diverse internet connectivity types, provided by diverse ISPs, running across redundant hardware that can provide the highest possible reliability. The big question has always been how to manage multiple links from multiple ISPs efficiently and effectively.

The answer can be found in a growing group of devices that are increasingly referred to as WAN Path Controllers. These devices accept multiple WAN connections of almost any type and then intelligently load balance. In some cases they can aggregate and bond bandwidth from multiple diverse connection types. They generally support fixed lines (Fibre, DSL, FTTC, EFM), cellular (3G/4G/LTE) and even satellite links, but their key focus is to make the monitoring, utilisation and management of those links as easy as possible, using whatever affordable connectivity is available.

WAN Path Controllers using multiple connection technologies, from multiple ISPs, offering affordable, reliable, and highly available internet connectivity anywhere: this is the future, reliability through diversity. NC