Get more out of your connections In essence, bonding means real aggregation of bandwidth of all WAN media to be bonded. Our bonding routers enable you to bond different networking media, such as DSL, 4G/LTE or cable, from multiple providers. The advantage of bonding internet lines compared to load balancing, is that with bonding a single virtual broadband line is created based on the combined broadband bandwidths, providing you with more bandwidth for your applications. A further advantage of bonding is the fact that an interruption of a bonded line will not affect the applications using the bonded link. If one link fails, others remain to ensure your Internet access.
As each connection requires its own particular bandwidth and latency characteristics, we ensure that autotuning and QoS settings are integrated, making it possible to provide connections that offer consistent service tailored to specific requirements. A further advantage of bonding is the fact that an interruption of a bonded line will not affect the applications using the bonded link. That is why bonding is the best solution for critical applications.
Another important aspect is the capability of prioritizing applications for business critical operations.
Technical explanation of bonding in practice The bonding router connects to the VPN Hub through a VPN tunnel. The VPN tunnels that connect the routers to Hubs are composed of channels. A channel is the name for a SSL-encrypted TCP connection between a module slot in a router and the WAN interface of a Hub. Hosts send packets upstream to the Node that routes traffic through their LAN. The Node fragments and encrypts these packets (and if relevant features are activated, compresses them), then transmits this traffic to the Hub over an appropriately aggregated set of channels in the tunnel. The Hub decrypts and reintegrates this traffic coming from the Node and transmits it to the open internet or other Nodes in the deployment. Incoming packets generally follow the same path in the opposite direction.
The VPN hub is located within M2MBlue’s own data center. As an alternative to a physical VPN Hub (bare metal setup), M2MBlue also offers a virtualized setup. Envisioned primarily for larger customer setups or carrier/telco requirements, this solution could be installed at the customer's premise (data center) if desired. The Virtual VPN Hub is a pure software application that works just like a physical hardware VPN Hub and can be operated as an Amazon AWS Cloud service or as VMware virtual machine on any existing server worldwide. Multiple instances are possible to support dedicated environments for each customer (type).
By default, M2MBlue terminates and collects the IP traffic to/from the public internet, safely secured by firewalls and web security gateways. Optionally, we can set-up a direct link (VPN or private) to customer’s data center to support end-to-end VPN connections.
But what is now the difference with load balancing? The two concepts are often mixed up. Chronological speaking, load balancing was the first of these two to be used in both enterprise and home networking environments. With load balancing, traffic is distributed across a few connections. Unlike bonding, these connections remain separate and you do not need a hub in a data center to bond these connections together. Because traffic is only running over one tunnel, it can’t take advantage of the additional bandwidth available on other active channels.
Both bonding and load balancing spread the network load on individual links while ensuring that the data is routed to the functional connections isolating the non-functional ones. The main difference lies in the level of the implementation of the function of sharing the burden of network data traffic.
Compared to bonding, load balancing has a few limitations: (1) the load balancer can only use the bandwidth from a single line for a single task plus you cannot combine the inbound bandwidth, (2) the balancing router has multiple broadband lines available to it. When the router uses all the bandwidth from a line, it switches new connections to the second lines. In case a load balanced connection fails, the session will be terminated and a new session has to be initiated which will be routed through a functional connection from the pool of the available connections. This means that if, for example, a local network user has started a Skype call and the used connection is interrupted, the session will be discontinued and the conversation has to be terminated. The one calling has to start a new session through another functional link. It is clear that this can’t be the case in a corporate environment.
We can conclude that load balancing can’t cope with current business connectivity needs.