(Editor’s note: Today’s post is by guest blogger and EMA Senior Analyst, Tracy Corbo)
One of the many insights we gained while compiling EMA’s recent WAN Optimization Controller (WOC) Radar Report, and after talking to numerous WOC users, is that there are many situations where dedicated WAN links are simply not an option. Traditional WOC solutions were designed to optimize expensive, dedicated WAN links. However, in some deployment cases, particularly those in remote geographical locations or developing countries, dedicated WAN links are not even a choice either because they are not available, are cost prohibitive, the workforce is mobile or the worksite is temporary. That means that these locations must settle for whatever network connectivity is available whether it is satellite, DSL, broadband, or even (brace yourselves) dial-up.
In our WOC Radar Report we came across a few vendors that are providing solutions that target these types of deployment scenarios – Ipanema, Talari, and Xtera. Each has a somewhat similar approach, but all are based on the assumption that they can make use of multiple external network access connections, whether they be traditional WAN (MPLS, Frame Relay, ISDN, etc.) or broadband Internet (DSL, cable, 3G/4G, etc.) to augment or replace expensive private dedicated WAN links and still deliver reliable, high bandwidth connectivity at a lower cost. I recently spoke with another vendor, Xroads Networks who could also be added to the list and I’m sure there are others.
But there is a dilemma for these vendors when it comes to getting the word out about their creative alternative approach, because it does not really fit neatly into any of the existing market taxonomies for, say, either WOC or Application Delivery Controllers (ADCs). Consequently, they tend to get the short end of the stick when they are lumped with well-established market leaders such as Riverbed and F5, because they do not offer the same feature sets as these other players. Ironically these solutions are different enough that they tend to be complimentary to those other solutions and can often be found deployed right alongside a WOC or ADC. Subsequently, these vendors (out of sheer frustration, if nothing else) try to find creative ways to define their market position and help educate customers on their unique approach to solving application performance issues. In the case of Talari and Ipanema, they have latched onto the term “WAN Virtualization,” but this does more to confuse the issue than create any unique market definition. This is not about classic virtualization on a VM nor network virtualization (a la VLANs) and it will only serve to further obfuscate the point rather than clarify it.
At the risk of oversimplifying the approach, what all these vendors do is bandwidth aggregation across multiple Internet or WAN links using algorithms to determine the optimal path for different traffic types while also providing redundancy. Each vendor has their own approach:
– Ipanema dubs theirs the Autonomic Networking System
– Talari calls their approach the Adaptive Private Networking (APN)
– Xroads uses simply Unified Bandwidth Management
– Xtera just discusses theirs as a combination of WAN load balancing and WAN packet shaping technology.
These are all appliance based solutions that can usually be deployed as complimentary solutions to existing WOC or ADC devices, providing enhanced application performance.
Those that offer aggregation of traditional Internet access links, such as DSL/Cable, also have another compelling aspect of their offering – drastically lower costs versus traditional WAN. In terms of $/Mb/month, mass-market Internet access is typically an order of magnitude less expensive than dedicated WAN services (i.e. tens of $/mo versus hundreds of $/mo). By aggregating two or three such consumer-oriented links, the load balancing and redundancy achieved means that reliability levels similar to dedicated WAN can be realized at a fraction of the monthly costs.
It is difficult to say whether this is an emerging market or subset of the existing markets. It might be time to take a closer look at this group of vendors and examine their feature/function sets and architectural approach in greater detail. While these solutions do not have the same mainstream acceptance as WOC and ADC products, they are worth taking a closer look at when network/application planning and operations teams are faced with application performance issues. There is always more than one way to tackle a problem and practitioners should be aware of all their options – this new alternative may by the only viable option and may also save a lot of money along the way.