But not every provider experienced similar levels of disruption. Like any network engineer, we wanted to know why. Our data held those answers.
The rules of traffic engineering are reset.
Network operators assess service requirements based on known or predicted traffic patterns. They take into account the location of users, their patterns of usage, and the type of services they access.
Predictions are typically based on foreseen events, like seasonal holidays, major sporting competitions or streaming game releases.
In pandemic times, some operators performed better than others at re-engineering around growing demand and handling the pressures they faced.
On the one hand, operators faced uncertainty as baseline usage expectations were reset. Any assertion of known and predicted traffic usage in the period was lost. This meant that any provisioning or changes required were reactive, potentially increasing the likelihood of them disrupting service availability.
Some operators did predict a path through this and started making changes to their networks in anticipation. As work-from-home and lockdown orders approached, they beefed up infrastructure and altered peering and load-balancing configurations.
With a critical mass of operators, all making changes came a precipitous rise in global network disruptions. The size and duration of the disruptions indicated they weren’t the result of capacity shortfalls, but rather maintenance windows brought about by intentional network state changes.
This may not have helped you then...
While these changes helped network and cloud operators weather huge user demand, Australian users may have been adversely impacted.
Maintenance windows are set with the intention of inconveniencing the least number of users. For hyper-scale and multinational operators, maintenance may be scheduled out-of-hours for the Americas and EMEA, where user bases are most extensive, but that often puts it inside office hours for A/NZ.
For A/NZ businesses, this has placed fresh importance on how they select their carriage and cloud service providers. They’re now looking much more closely at the characteristics of connections and how it affects them.
In times like these, a domestic provider may seem an obvious choice over a global one, if only because their network maintenance (and therefore disruption) windows are more likely to align to local needs. But their upstream providers and peers also need to be factored in: where are the upstreams located? What impact does work they undertake or disruptions they experience have on the domestic provider and therefore flow through to the service you consume?
Understanding these kinds of specific operator practices is key to ensuring your business can plan and effectively communicate with stakeholders and vendors to ensure minimal business disruption even when unforeseen traffic shifts require providers to make more frequent network state changes.
The last few years’ rapid acceleration of cloud adoption and widespread use of SaaS applications has brought with it increased deployment of next-level network intelligence technologies that are capable of bridging the gap between enterprise networks and external ones like public cloud and Internet service providers. Only with that level of insight and visibility will you be able to develop your own benchmark performance to compare providers and performance levels effectively.
… But it may wind up helping you now
Once you have that visibility, you can start to use it as a foundation to negotiate performance-based service level agreements (SLAs) with carriage and cloud service providers.
SLAs drive behaviour. The past six months have put SLAs in focus and exposed a need to derive a more dynamic SLA structure that provides something of a win-win relationship for all parties involved.
For many providers, SLAs are becoming an important differentiator in their offering, by proactively rewriting customer contracts around user experience guarantees. Users, for example, may be contractually guaranteed certain performance levels, like the ability to log in and load the first page of a business-critical application in X seconds. This is enabled by the ISP working with the (hosted) application provider to optimise traffic routing and experience delivery.
SLAs are a two-way street, however, and as business customers leverage new technologies to gain more visibility into the performance of external connectivity providers, they will find themselves in a better position to compare providers and negotiate positions, with data to support their stance.
With providers and their customers alike having found an operating rhythm, we expect more of these types of conversations to now occur.