Network Support Services and Infrastructure in Canada

Everything between the wall socket and the application, which is why the term is vague. In practice network support services means monitoring and alerting on the links and devices, configuration management and backups, firmware and patching, capacity and performance work, and the design changes that stop the same incident recurring. Network infrastructure services is the broader term that adds the build: the data centre, the hypervisor, the load balancers. We do both, and the assessment tells you which you're actually short of.

Five groupings cover almost everything a business buys. Connectivity, which is the circuits and the SD-WAN over them. Switching and routing inside your sites. Security, meaning segmentation, firewalls and the perimeter. Application delivery, which is load balancing and failover. And the support wrapper: monitoring, patching, documentation and somebody to call. Most vendors sell the first three well and thin out on the last two, which is where outages actually come from.

Work with it, in most engagements. A good internal team knows the business constraints no outside network services provider will learn in six weeks, and what they usually lack is a second CCIE for the hard design calls and cover at 3 a.m. We take the design, the cutover and the out-of-hours, they keep the day-to-day and the relationships. Where there's no team at all we'll run the network support services outright, and we'll say which arrangement fits.

Usually, and not always, and the honest answer needs your traffic. SD-WAN wins when most of your traffic is heading to the internet or a cloud provider rather than back to a head office, because then you're paying MPLS prices to haul packets the wrong way. It wins less when you have hard latency requirements between sites or a regulator with an opinion about your paths. We model it against your actual flows before quoting anything.

Ask what the workload does, not what the industry is doing. Steady, predictable, licence-heavy workloads are often cheaper on your own hardware, and hyperconverged has narrowed the operational gap a lot. Bursty, seasonal or new workloads usually belong in a cloud. Most Canadian mid-market estates we assess end up split, and the useful output is a per-workload decision rather than a strategy slide. Data residency requirements sometimes make the call for you.

Vendor-neutral, and the certifications happen to be Cisco because that's where the deep enterprise routing and data centre credentials live. In practice we design and run Cisco, VMware, F5, Palo Alto and Fortinet estates, and network support services from us don't come with a resale quote attached. If the right answer is keeping what you have for another three years, that's the recommendation you'll get.