[wp_tech_share]

But Market Growth is at Risk Due to Service Confusion

In 1880, Thomas Edison said “We will make electricity so cheap that only the rich will burn candles” and today, no enterprise can operate without power. Since the upheaval of a global pandemic, IT manufacturers are seeing Edison’s promise in a new light: converting Wireless LAN into a utility will accelerate enterprise productivity. Along the way, manufacturers stand to gain a steady stream of high-margin, recurring revenue.

Manufacturers of Campus LAN equipment (enterprise-class WLAN and Switching) generated $29B of revenue in 2022 with an anticipated 5-year CAGR of 2%. New entrants believe the market is ripe for an innovative service offer and incumbents are looking to accelerate market growth.

The Vision is Compelling, but its Instantiation Remains Unclear

Industry visionaries have therefore married the vision of “Wi-Fi as a utility” with the cloud-enabled “As-a-Service” (aaS) technology abstraction to create a compelling proposition – but it is rare to find two manufacturers who describe the service in the same way.  With each variation having different commercial, financial, and technological implications, it’s a complex landscape for IT departments to navigate.

And so, the killer question remains unanswered:  Will the Enterprise Campus NaaS opportunity succeed in expanding the campus network market? Or will it be relegated to a differentiating feature of existing solutions, fueling competition without growing demand?

New entrants to the Enterprise Campus NaaS market believe the former proposition is true.  They have developed cutting-edge technology and innovative commercial strategies to meet enterprises’ needs.  Meanwhile, incumbent vendors’ are positioning services that have deep feature sets, well-developed channels, and strong brand awareness.

A common framework for defining Enterprise Campus NaaS is critical for manufacturers to quantify the market opportunity and hone their strategies.  For the service to flourish, their customers’ IT departments must be able to understand and compare the available services.

Defining the Cloud Consumption Model in a Campus Context

The three words “As A Service”, or aaS, have become synonymous with the cloud computing model.  The aaS extension is appended to different words (e.g. Infrastructure, Platform, Software) to denote different levels of technology abstraction.  When applied to campus network IT, the cloud consumption mode is often called NaaS, or Network As-a-Service. This term lends confusion as it can also denote Wide Area Network services.

Since terminology matters when new markets are being developed, we define Enterprise Campus NaaS as the delivery of campus connectivity at Layer 2 and Layer 3 of the OSI model (such as WLAN and switching) within the premises of an enterprise or organization by means of a service that adheres –at least partially– to the cloud consumption model.

To provide a common framework for comparing the offers on the market, we define the four key parameters of the cloud consumption model below and explore how these can be instantiated in campus networks.

  1. Cloud Consumption Services Are Outcome-Oriented

With an outcome-oriented service, an enterprise no longer purchases technology. Instead, it purchases a service based on a result it expects to attain. For example, the Open Data Center Alliance has defined IaaS outcomes such as millions of IO operations, or GBs of disk capacity.

In a campus networking environment, an enterprise could purchase a solution defined by a consistent Wi-Fi signal level over a given area or a minimum download speed for a specified number of devices. However, due to the complexity of these models, many Enterprise Campus NaaS providers opt to structure their service around the underlying technology, charging a fee based on the number of APs or ports.

A truly outcome-based Enterprise Campus NaaS must be accompanied by an enforceable Service Level Agreement, which remains another impediment. Manufacturers are well aware of the challenges and costs associated with implementing SLAs.

  1. The Cloud Consumption Model is Elastic

An elastic “aaS” offer appears infinite. The purchaser of PaaS is not bound by the fixed dimensions of a server or hard drive; the service provider ensures the capacity and desired reliability are available.

For a university whose students have just discovered the latest bandwidth-hogging application, an elastic Enterprise Campus NaaS would absorb the unexpected traffic peaks with no costly design changes or additional hardware.

However, the tight coupling between WLAN hardware and its physical installation represents a challenge for service providers. Some of the offers on the market are designed to have a certain elastic nature, but their upper limits will remain constrained by the on-premises hardware installed.

  1. Cloud Consumption Services have a Recurring Price Structure

In its simplest form, Enterprise Campus NaaS, is priced with a subscription fee:  a conversion of capital cost (of APs and switches) to a recurring, operational expense.

This can be attractive for a distributed retail operation opening a new store, whose large, up-front cost of the network infrastructure disappears. The new location would increase the company’s IT bill in the same proportion as existing locations, simplifying cost attribution and recovery.  In these types of service offers, the cost of financing the hardware is often blended into the monthly price.

For “aaS” offers that are also elastic and outcome-oriented, another commercial structure becomes a possibility: consumption-based pricing.  With this model, a university network that benefited from an Enterprise Campus NaaS would have hardware in place for near-infinite usage, but the monthly bill would dip down in the quiet summer months.

Given the large cost of network hardware on premise, manufacturers may find it difficult to charge true consumption-based pricing.  Enterprises will have to commit to prescribed contract lengths or minimum monthly charges, even if the service price varies somewhat according to usage.

  1. Cloud Consumption Services are Maintenance-Free

Whether it’s IaaS, PaaS, or SaaS, hardware maintenance is performed by the service provider. For Enterprise Campus NaaS, maintenance, or life-cycle, services are the blurriest of the cloud consumption parameters.

Traditionally, campus IT manufacturers have shied away from delivering ongoing life-cycle services, avoiding direct competition with MSPs, their valued channel partners. However, last year Home Depot announced it was outsourcing portions of its campus network operations to HPE, whose executives have been emphasizing the long-term profitability of “aaS” offers. In contrast, Juniper Mist and Cambium Network’s Enterprise Campus NaaS announcements focus on enabling their channel partners.

Whereas new entrant Shasta Cloud is promising to revolutionize the way MSPs deliver Enterprise Campus NaaS to their clients, startups Nile and Meter are focused on delivering the full gamut of life-cycle services directly, as well as via channel partners.

The most difficult phase of the technology life-cycle to include in Enterprise Campus NaaS is the hardware end-of-life. A truly “evergreen” service, would include hardware upgrades as the 802.11 standards evolve –without an enterprise paying for them outright– but Enterprise Campus NaaS is still too new to have put this phase to the test.

Related Blog: The Role of Ethernet in Wi-Fi as a Utility (written by David Rodgers, EXFO; published by Ethernet Alliance)

 

How is the industry positioning Enterprise Campus NaaS?

Over the past 6 months, we have interviewed over a dozen industry participants who have commercialized, or are planning on commercializing, some type of Enterprise Campus NaaS.  A minority of the services met all of the cloud consumption characteristics.  However, all of the services met at least one criteria, with subscription-based pricing being the most popular.  Fewer than a third of the Enterprise Campus NaaS offers had some form of elasticity or contained an evergreen provision to upgrade hardware at no additional cost.

With industry players approaching Enterprise Campus NaaS from different angles, it follows that their commercial strategies vary, with a focus on different customer verticals, segments and channels.  However, there are three broad categories of Enterprise Campus NaaS that are beginning to emerge.  As the market matures we expect to see vendor strategies consolidate according to which of the three types of Enterprise Campus NaaS they envision: Enabler, Turnkey, or Wi-Fi as a Utility, as depicted below.

Those who are navigating the complex Enterprise Campus NaaS landscape can look to history, at the evolution of the electrical power grid. Thomas Edison chose to back a power distribution system based on Direct Current —convinced that Alternating Current transmission systems were too dangerous to the public. The competing models battled it out for over a decade before Edison Electric merged to form General Electric, and AC distribution became the worldwide standard.

Enterprise Campus NaaS brings the allure of easy-to-manage, ubiquitous WLAN at a time when businesses depend on wireless connectivity more than ever. However, to effectively market the service, the industry needs to converge on some common definitions. Once that happens, enterprises will be able to take their Wi-Fi coverage for granted, thinking about it as much as they think about electricity — about once a quarter, when they pay their bill.

More information on the different approaches to Enterprise Campus NaaS and a quantitative analysis of the market will be included in the advanced research report entitled Campus NaaS and Public Cloud-Managed LAN, to be released in June 2023.

Related Blog: The Role of Ethernet in Wi-Fi as a Utility (written by David Rodgers, EXFO; published by Ethernet Alliance)

[wp_tech_share]

2022 marked a record year for broadband spending around the world, as service providers forged ahead with major network upgrades and expansions. In many cases, the focus of these investments was to expand the reach of fiber for business and residential services with the ultimate aim of de-commissioning legacy copper and DSL networks.

A major part of these network upgrades was the investment in PON technologies with the ability to deliver 10Gbps of bandwidth across a single OLT port, which is then shared by dozens of subscribers. The short-term goal is to be able to deliver symmetric 1-5Gbps of bandwidth consistently to each residential subscriber. The focus on delivering these speed tiers has resulted in a significant jump in the purchasing of 10Gbps-capable technologies, including 10G EPON, XG-PON, and XGS-PON. From 2020 to 2022, spending on OLT platforms and ONTs supporting 10Gbps technologies jumped 308% (Figure 1).

Figure 1: Worldwide 10Gbps-Capable PON Equipment Revenue

 

While these technologies will serve most operators well for the next 5 years operators in a growing number of markets want to ensure that the significant investments they are making today in expanding their fiber networks and ODN won’t be regrettable investments and that there is a technology roadmap in place that keeps them ahead of their competition in terms of speeds and latency, but also allows them to achieve a number of architectural goals, including delivering both residential and enterprise services using the same technology and ODN; collapsing access and aggregation networks to reduce the total number of network platforms; providing a simplified upgrade path through co-existence of multiple PON technologies, and; delivering wholesale mobile transport services.

Bandwidth demands show no signs of slowing, with the ITU having estimated that worldwide bandwidth consumption grew at a compound annual growth rate (CAGR) of 50% from 2015 to 2021, reaching a total of 932 Tbps, up from 719 Tbps in 2020. With governments and operators alike focused on expanding their networks to get more homes and businesses connected, as well as applications like virtual reality (VR) and online gaming set to expand, bandwidth consumption will almost certainly accelerate throughout the remainder of this decade.

For some operators in highly-competitive environments, 25G PON is the appropriate next step, as short-term demands for bandwidth beyond 10Gbps and for the need to address both residential and business customers from a single ODN push them to act within the next 1-2 years.

Meanwhile, the ITU-T’s 50G PON Standard and corresponding prototype platforms and components continue to evolve quickly, as operators and equipment vendors look to accelerate the availability of products so that they can undergo the rigorous testing and homologation required of any new technology. Considerable effort has already gone into defining the physical layer parameters, latency requirements, and Forward Error Correction (FEC), among other elements.

Already, a number of operators have either conducted early lab trials of prototype equipment or have endorsed the technology as their next choice, including China Mobile, China Telecom, China Unicom, Globe Telecom, Orange, Saudi Telecom, Swisscom, Telefonica Spain, Telekom Malaysia, and Turkcell. Other operators are keeping an eye on the standardization process but are also largely focused on their current rollouts of XGS-PON to make any formal commitment beyond that.

Additionally, a component ecosystem is emerging quickly, driven largely by system vendors who want to get products to market quickly, as operators want to be absolutely certain that the power budget requirements and dispersion penalty, along with the use of digital signal processors (DSPs) does not force any change in the existing ODN.

 

Recent Steps Forward

A number of major steps forward for 50G PON were announced in September 2022, during the ITU-T Study Group 15’s Plenary Meeting. The most significant announcement was the agreement on details for the simultaneous coexistence of all three ITU PON technologies (50G PON, XGS-PON, GPON) on a single ODN. Previously, simultaneous coexistence with GPON and XGS-PON had not been defined, meaning that operators would have to upgrade their GPON networks to XGS-PON prior to beginning their 50G PON deployment.

With the addition of a third upstream wavelength band (1284-1288nm) to the G.9804.1 standard’s existing 1260-1280 and 1290-1310 bands, 50G PON, XGS-PON, and GPON can now live together on a shared ODN. Additionally, combo PON implementations can now be supported using 50G + XGS-PON, 50G + GPON, as well as all three modes (50G + XGS-PON + GPON).

The support of both simultaneous coexistence as well as combo PON implementations is critical addition as operators have said time and again that they do not want to disrupt their ODNs when moving to a new technology. Additionally, operators are expected to make their transition to 50G PON through the use of combo PON, which takes advantage of the existing space in the central office, requires no modifications to the ODN, and does not require the use of a WDM multiplexing device, which can result in optical power loss.

 

Challenges Remain

50G PON represents a significant improvement in bandwidth availability and latency over today’s 10Gbps technologies. However, these benefits don’t come without their challenges. The biggest technical challenges are in the PHY layer. Specifically, the optical power budget required, dispersion penalty, and intersymbol interference (ISI) are all potential hazards in 50G PON systems. As bandwidth increases, overall performance typically declines, especially when the existing ODN defines a 32dB power budget. The use of DSP technology can reduce or eliminate these PHY layer issues. However, previous PON technologies did not use DSPs, so operators will want to test this thoroughly and ensure that point-to-multipoint communications between the OLT port and ONTs are occurring as expected and without error. The DSPs specifically help to reduce the dispersion and bandwidth limitation penalty, as well ensuring that lower-bandwidth GPON and XGS-PON ONTs are supported more efficiently.

At this point, current 50G prototypes are asymmetric, delivering 50G downstream and either 25G or 12.5G upstream. Though system vendors are working through the best options for delivering consistent, symmetric speeds and have already delivered some prototypes using semiconductor optical amplifiers (SOA) and FPGA-based DSPs, the ITU-T SG15 agreed back in September 2022 to further study the options for delivering symmetric speeds. Clearly, operators would prefer a symmetric option as early as possible. But the dramatic increase in downstream bandwidth and billboard speeds should more than suffice until the upstream technologies and components have been standardized and implemented in OLTs and ONTs.

 

Opportunities Continue to Grow

Though early, Dell’Oro Group believes total 50G-PON equipment revenue will increase from less than $3M in 2023 to $1.5B in 2027. Much more significant growth is expected after 2027, as operators begin to evolve their 10Gbps PON networks to next-generation technologies (Figure 2).

Beyond being able to anticipate future bandwidth growth coming from consumer applications such as VR, AR, online gaming, videoconferencing, and 8k video, 50G PON positions operators to address business services. Specifically, 50G PON allows a provider to offer four 10G Ethernet connections, split among multiple businesses. Additionally, 50G PON is ideal for POL (Passive Optical LAN) deployments, where fiber can be run to the desktop and deliver connectivity with less power, rack space, and less cooling than traditional point-to-point Ethernet architectures.

Figure 2: Worldwide 50Gbps PON Equipment Revenue

Figure 2: Worldwide 50Gbps PON Equipment Revenue

 

Similarly, 50G PON has applications in the backhaul of public Wi-Fi hotspots as well as private wireless LANs, both of which will see significant bandwidth growth with the availability and deployment of Wi-Fi 6E and Wi-Fi 7. Wi-Fi 6E allows individual subscribers to burst to 9.6Gbps while Wi-Fi 7 quadruples that throughput to nearly 40Gbps. Additionally, the Wi-Fi 7 standard defines extremely low levels of latency and jitter, which the evolving 50G PON standard is also incorporating.

Finally, as operators continue to converge their residential, business, and wholesale fiber networks onto a single ODN, 50G PON is envisioned as the universal technology to deliver services across those networks. Mobile midhaul and fronthaul applications, expanding IoT devices and services, wholesale fiber access to macrocells, backhaul of fixed wireless access (FWA) nodes—all of these can in theory be delivered using 50G PON. Other applications and use cases are certain to emerge as operators continue to reap the benefits of converting their disparate networks onto a shared ODN, with throughput and services delivered via 50G PON.