In a previous life as a product manager at service providers GTE and Verizon, I discussed SLAs for latency of services like frame relay and MPLS thousands of time. The focus on SLAs is always tied to the critical business services – meaning the transport latency needed to meet the needs of the applications and end-user requirements.
A recent article by Shamus McGillicuddy from TechTarget brought to light an interesting concept – measuring Human Latency. Human latency is typically defined as the amount of time it takes to handle normal network tasks like turning up a port, reassigning a VLAN or creating a new network/subnet. While network latency is normally measured in milliseconds, the impact of human delays is typically measured in hours, days or weeks.
The network has seen a paradigm shift. When networks were more static, IT teams could get away with higher human latency, but as the proliferation of devices explodes and the spike in virtualization continues, businesses are negatively impacted with human latency. When end users sit around waiting for a port to be activated or access to a VLAN, they can’t do their job.
Before, network engineers spent a lot of time and effort focusing on network latency, but didn’t put a lot of emphasis on human delays. A perfect proof point is asking your network team to define their SLAs for network latency and they will typically say “our provider guarantees latency no more than xx milliseconds.”
And then ask the SLA for normal changes and the answer is generally “As quickly as we can” or “hopefully the same day.”
Even for “simple” changes like activating a port or reassigning a VLAN, it normally comes through the help desk but typically requires a network engineer expert to make the change via CLI. While the change itself should take less than 60 seconds to complete, the time grows to hours and days because of several factors:
- The CCIE or network engineer is busy with other tasks
- The first tier help desk doesn’t have access to make the change
- Multiple handoffs across many silos
Many organizations give lip service to automation and pushing tasks to less experienced staff, but haven’t implemented a lot because of security and human error concerns. One of the goals of Infoblox is providing practical network automation. Our Automation Task Board leverages an intuitive interface to help deal with these common tasks in a matter of seconds, not hours. More importantly, the solution provides the safeguards to enable lower level staff to make common changes but since it’s pre-approved and tested by the network engineers, there is comfort the tasks will be done correctly.
So the next time you discuss SLAs or the amount of time it takes to implement business services, don’t forget to measure the impact of human latency. If you’re using a sundial instead of a stopwatch to measure how long it takes to make typical changes, it’s time to think about how network automation can help you.