Redundancy, not SLAs
Business grade data communications services come with Service Level Agreements (SLA), but SLAs are a joke. As an example, Verizon promises 100% availability but they back up their SLA with this commitment:
For each cumulative hour of Network Unavailability or fraction thereof in any calendar month, Customer’s account shall be credited for the pro-rated charges for one day of the Verizon monthly recurring charge.
So if your business data services cost $500/month, Verizon will credit you less than $17 for each hour their service is down. What’s your cost in lost productivity? In being unable to process credit cards? For many businesses the answer is hundreds or thousands of dollars per hour, or more.
Redundancy is the best solution
With two independent sources of Internet, when one service fails you’re still connected. Only if both services fail at the same time do you lose Internet connectivity. With redundancy, your effective connection is more reliable than the better of the two services!
Of course there’s a catch. The services must be “independent” meaning a failure of one is not tied to a failure of the other. In real life, this means the two services must come from different vendors via different paths.Having services that use different strands in the same cable bundle doesn’t help if that cable is cut.
Cost
A more significant issue may be cost. Internet access can be costly, especially for business class services, and these are monthly recurring charges. Until now, if you needed reliable internet service, you just had to bear the additional cost of a second service.
netBlazr changes everything. Monthly costs can be zero! With netBlazr, there is a one-time cost to join the network but thereafter, basic service is free and premium services are available at a small fraction of traditional costs.
Implementing redundancy
The most common approach uses the dual wide area network (WAN) capability built into most modern firewall appliances. For example, the SonicWALL Pro 2400 security appliance includes dual WAN connections with a choice of active-active load sharing or fail-over configurations. As an alternative, there are separate dual WAN routers available from Cisco and others for less than $150.
Calculating reliability – for those who love math
We actually use the formula for calculating the probability of system failure given failure rates for each of several replicated services that make up the system. The formula is:
P = p1 * p2 * p3 * …
Where P is the probably of system failure and pi is the probability of the ith service failing.
This formula assumes independence of failure events. That means that the probability of service B failing, given that service A has already failed, is the same as that of B failing when A has not failed.
An example:
Let’s say your first service, which is advertised as “99.99%” reliable, is really only 99.95% reliable, i.e. the service is not working for 4-5 hours each year. Now you add a second service with only 99% reliability (i.e., it’s out of service 1% of the time) but you configure a reliable dual WAN router for redundancy. Now the combined failure rate is:
0.05% * 1% = 0.0005%
A 0.0005% failure rate means the combined service is available 99.9995% of the time. That’s better availability than provided by any single service offered by any Internet service provider. And with netBlazr, the incremental cost is tiny.