What to Expect when Signing up for NaaS in the Campus LAN
Campus NaaS is poorly defined in the industry, leading to market confusion. In this series of blogs, Siân Morgan explores the differences and similarities of the offers on the market and proposes a set of definitions to help enterprises and vendors speak the same language.
In 1999, lack of common definitions among industry participants had catastrophic implications to space exploration. The Mars Climate Orbiter, an interplanetary weather satellite designed to study the Martian climate, was caught in a fundamental miscommunication. While the team that developed the software calculating the thruster impulse was using pound-force seconds, the team that developed the software to calculate the craft’s trajectory interpreted the data as newton-seconds. Every adjustment of the spacecraft’s trajectory was off by 4.45. When the Orbiter reached Mars, it was 105 miles closer to the surface than expected, and in all likelihood, the $125 million craft was burned to smithereens.
While there may be similar challenges in understanding the different aaS or “as a service” offers available for LAN connectivity, the misunderstandings will be less explosive. Vendors such as Extreme, HPE, Join Digital, Meter, Nile, RUCKUS, and Shasta Cloud have aaS offers for the LAN that differ considerably, but we can start to make sense of the landscape by focusing on the common thread: a service model inspired by cloud computing. Figure 1 depicts the fundamental service and implementation characteristics of a cloud-computing-inspired offer.
Whereas terms like IaaS and SaaS have been around long enough for industry participants to understand the broad scope of the offer, when describing a layer-2 network used to connect PCs, mobile phones, printers, smart blinds, sensors, and the myriad other IoT devices within an enterprise’s campus, using the term NaaS usually leaves more questions than answers.
Appending a “C” to the acronym (CNaaS or Campus Network-as-a-Service) designates that the service pertains to campus connectivity, meaning connectivity within an enterprise or branch office. Figure 2 depicts where CNaaS lies within the network, as compared to the more traditional NaaS, IaaS, PaaS, or SaaS offers.
CNaaS Characteristics
Unlike a cloud-computing offer such as SaaS, campus networks require a large amount of equipment to be installed on-site and dedicated to a physical space. Radio frequency waves and cables are subject to the laws of physics, and WLAN APs can only be shared between enterprises if the physical office space is also shared. This limits two of the cloud-inspired implementation characteristics: most of the hardware cannot be shared, nor can it be centralized. CNaaS vendors have thus turned to other approaches to simulate aaS characteristics.
Among the CNaaS offers on the market, the cloud-computing characteristics have been instantiated as follows:
Outcome Oriented
An outcome-oriented service is priced based on an expected result, rather than on the number of ports, APs, or other technology deployed. For example, some CNaaS vendors, for instance Meter and Nile, offer service based on a combination of the number of devices, the floor space, and/or a series of user-experience metrics.
By purchasing a service based on an outcome, customers are effectively transferring responsibility for the underlying technology design to the provider of the service. Some vendors provide service credits to enterprises whose networks did not meet agreed-upon service level metrics. For enterprises with complex networks, the advantage of outsourcing the network design comes with an additional complexity of capturing the SLA in a contract.
Elastic
Since the maximum capacity of a LAN is determined by the hardware installed onsite, it is impossible to deliver a CNaaS offer that provides unlimited capacity for expansion. Some vendors, for example Extreme, provide rebates to enterprises to turn off ports or APs during quiet periods of the year. However, in this case, the hardware remains in place and unused and the cost to deliver the service is not altered significantly.
CNaaS offers, such as HPE’s Turnkey CNaaS, can also simulate elasticity by blending the price of the solution for a multi-site organization, allocating the bill for the service according to the network size at different sites. For example, if an enterprise has two branches, and Branch A has twice as many network users as Branch B, Branch A would pay two-thirds of the enterprise’s CNaaS bill. This distribution allows for better cost allocation within the organization.
Opex Price Structure
By centralizing and sharing computing resources, traditional aaS providers allocate abstracted portions of technology to each customer and charge the customer based on how much capacity it uses. Because of the limited ability to centralize and share the infrastructure of a campus network, this approach cannot be used with CNaaS offers.
To deliver an opex price structure, some CNaaS providers retain ownership of the hardware, with the consequence of a growing balance sheet. Other providers make use of a third-party financing company. Large vendors, such as HPE, can rely on established financing divisions to back the service. In any of these situations, financing adds to the service cost. In addition, most CNaaS offers come with a prescribed contract duration to ensure the vendor covers its upfront investment. In the few cases that offers do not come with a set contract length, one of the parties involved is assuming the risk of early termination.
Maintenance-free
Traditional cloud-computing services make use of centralized computing resources that are housed and maintained by the supplier, thus relieving the customer of installing, maintaining, and refreshing servers. The large amount of on-site hardware required for the LAN makes this maintenance-free characteristic difficult to implement, and CNaaS vendors have circumvented this with two different approaches.
First, many vendors partner with MSPs to deliver lifecycle services as part of the bundled recurring CNaaS fee. This is the case with RUCKUS and Extreme’s CNaaS offers that are delivered with the companies’ channel partners. Lifecycle services go beyond the 24×7 hardware support that is usually bundled with an equipment’s license. They may include network design, installation, and ongoing monitoring, while allowing an enterprise to apply and manage its own network policies and some local configurations.
In the second approach, CNaaS offers include 24×7 equipment monitoring by the vendor (such as Nile or Meter), as opposed to by the MSP. Monitoring is performed from a centralized location, using as much automation as possible to reduce costs. The degree of customer visibility and control over the network is a challenge for CNaaS vendors. Whereas some enterprises may wish for more of both, the cost of developing additional features to enable enterprise control undermines the maintenance-free concept.
In truly opex-structured, outcome-oriented, maintenance-free offers, hardware replacements are also included as part of the service. For instance, the CNaaS vendor would upgrade a network from Wi-Fi 6 to 6E and to Wi-Fi 7 at no additional charge to the enterprise. The events that trigger such a hardware upgrade are not always clearly laid out, highlighting the contractual complexity of this approach.
Enterprises considering an aaS construct for their LAN service should decide how important each of these cloud-inspired characteristics is to them and insist that their prospective service providers clearly define how each is to be delivered. As I will explore in my next blog, service providers may use the same terms, but the services vary depending on the providers’ business objectives.
