We all talk about Service Level Agreements when evaluating carriers that provide MPLS networks or telecom services. But what really matters to you?
The most common SLA categories relate to latency, packet loss, jitter and availability. But what is really important to your company? Would an SLA for application performance make sense to you? Of course it would! It takes negotiation abilities, which consultants like SD-WAN-Experts can provide. One must recognize that asking for more stringent SLAs exposes the carrier to more risk, so there may be a price to consider for these extra protections.
Common SLA (Service Level Agreement) Categories offered:
Latency: the round-trip ping time on a circuit. The quoted Latency SLA is typically 15% – 20% slower than actual performance, allowing the carrier network performance to drop without a penalty. When evaluating a network, the actual latency can often be tested and built into a custom SLA if your deal is big enough. Note that some Latency SLAs address the average backbone latency for geographic regions, not necessarily your particular network paths.
Packet Loss or Packet Delivery: this SLA item is self explanatory. If you have a private network, you expect near zero packet loss. Again, some carriers base the packet loss SLA on their backbone network, not your end-to-end network. Make sure you understand what is being offered.
Jitter: the jitter SLA measures delivery of packets in proper order. This is of particular importance to VoIP and Video.
Service Availability: Every network will have occasional failures, most commonly due to a local loop failure. The carrier needs and SLA with their local loop provider to assure that repairs are made promptly. A common SLA is four hours to recover from a failure. But different carriers start the clock at different times. For a managed network, the clock might stop when the second test ping does not respond from the NOC. Some carriers do not start the clock until the customer initiates a trouble ticket, which might be hours after the start of a failure, depending on the time of day and global expanse of your network. This particular measure is of limited value if you experience large numbers of brief outages that do not add up to a large number, but are disruptive to your operations.
Service Installation: this is a guarantee of how long it will take to install a circuit at each specified location on your network. But you must understand the fine print. Many carriers will credit the non-recurring charge for an installation that does not meet the SLA. But many contracts have the non-recurring charge (NRC) waived, so the credit would be zero!
Proactive Outage Notification: this metric is found with only some carriers. It defines that fact that the carrier will inform the customer of an outage automatically, without the customer needing to open a ticket.
Fast Fail-Over: measures the speed of failover from a primary network path to secondary network path in a failure situation.
Do You Have the Measurement Tools? If you (and your carrier) have the tools and your network would suffer greatly from rising packet loss of latency, consider negotiating an SLA that samples latency and packet loss on specific paths of your network every minute or some other agreed upon period. Then set a threshold of how acceptable values and how often they can occur before requiring and SLA credit. This approach helps address issues that are intermittent enough that they don’t affect the monthly numbers, but for you, the customer, they interrupt your operations. You need to have a substantial network for a carrier to even consider this level of granularity in an SLA.
When all is said in done, why not have an SLA that specifies the responsiveness of your applications? Or maybe your general satisfaction with the network and its support staff? Boilerplate Service Level Agreements are subject to negotiation if your network is substantial enough. For all the service level credits that might be offered, none of them will truly compensate you for service interruptions which affect your operations. But remember, it is your job to assure that your network has redundancy so that a carrier service interruption at one location does not stop your network from operating. Perhaps performance may not be as robust on the alternate path, but you need to plan for failures and utilize the alternate solution that is appropriate for the type of office you are servicing. The experts at SASE Experts can assist you in designing a solid network for your company. Give us a call!
