Working out at the gym. Waiting in the doctor's office. Shopping in the grocery aisle. Meeting in the conference room. With digital transformation, these types of activities are increasingly now hybrid, with many virtual options. At the same time, the demand for insights with AI/ML applications are growing, from generative AI and chatbots to medical diagnostics/treatment and fraud detection.

The rising use of online applications and analytics is generating large amounts of data that need to be moved swiftly, and as a result, users and devices are demanding more bandwidth. According to GSMA, 5G connections will grow to 5 billion by 2030. Analysys Mason forecasts that there will be 6.2 billion fixed and mobile connected IoT devices by 2030, up from nearly 1.8 billion at the end of 2020.

Adoption of 1G+ broadband also continues to grow rapidly. Based on the latest OpenVault Broadband Insights Report, average per-subscriber broadband consumption approached a new high of nearly 600 GB per month at the end of 2022 and the percentage of subscribers provisioned for gigabit speeds more than doubled Y/Y to 26%. What's even more interesting is that the percentage of power users consuming 1TB or more per month was 18.7% Y/Y, and "super power users" consuming 2TB or more per month grew 25% Y/Y in Q4CY22.

Analysys Mason forecasts global fixed internet and cellular data volumes to rise to a combined total of 18.5 zettabytes (one zettabyte = one trillion gigabytes) worldwide by 2028 -nearly 3 times what it was in 2022.

Network Implications

What does this all mean? High-speed broadband and 5G mobile access are enabling users to consume more bandwidth, and seem to be driving "induced demand", where, in this case, increasing the bandwidth supply can create more demand.

In particular, video is highly bandwidth-intensive and continues to dominate traffic patterns, whether for entertainment or real-time communications.  For example, depending on the quality, short-form videos can add up to 300MB to 800MB per hour, a videoconference call can consume from 800MB to 2G/hour and streaming video can generate 2G to 7GB/hour.

Given these traffic rates, service providers and cloud operators are looking to scale for today and the future to keep up with user demands. Delivering high-quality user experiences is important for providers, and relies on a network infrastructure that can have the capacity and control to provide high-quality services.

Growing network capacity can require adding more line cards to modular routing systems as well as more routers, which can drive up complexity and space consumption with more hardware expansion. For example, scaling to 230T aggregate throughput using 115.2T modular platforms could require up to six systems, which is estimated to be nearly 80 kW power consumption^[1].

What if you could double the performance of your phone, without replacing it entirely? At Cisco, we have made investments to help scale routers without complete replacement or sacrificing simplicity and operational efficiency.

New Cisco 800G Innovations

With market-leading densities and space efficiency through the industry's first 28.8T line card powered by the Silicon One P100 ASIC, we are introducing 800G capability to the modular Cisco 8000 Series Router, which can scale to 230T in a 16 RU form factor with the 8-slot Cisco 8808, and up to 518T in the 18-slot chassis (see press release).  At up to 15T/RU, we estimate that our dense core and spine solutions can deliver industry-leading bandwidth capacity and space savings, with up to double the capacity of competing single chassis platforms and up to 6x more space efficient compared to distributed chassis solutions.

These new line cards can support 36xQSFP-DD800 ports, which can enable the use of 2x400G and 8x100G breakout optics, and deliver market-leading densities with 72x400G ports or 288x100G ports per slot. The reason we can double the density is because the P100 uses state-of-the-art 100G SerDes technology that can achieve higher bandwidth speeds in the same footprint.

Instead of six 400G modular systems, one 800G 8-slot modular system can achieve 230T with up to 83% space savings, up to 68% energy savings or ~215,838 kg CO2e/year ~GHG savings. To put it in perspective, these carbon savings are the equivalent of recycling 115 tons of waste a year instead of going into landfills^[2].

In addition to sustainability and operational cost benefits, our customers can also protect their pluggable optics investments since Cisco QSFP-DD 800G can support backward compatibility to lower-speed QSFP-DD and QSFP modules.

Operational Simplicity

Doubling the density in the same footprint can also mean less hardware to manage, which can help simplify operations. Managing traffic with a high-speed network might seem challenging, so we're also providing more visibility, granular and scalable services health monitoring, closed-loop network optimization and faster provisioning with Cisco Crosswork Network Automation . These capabilities help customers consistently meet SLAs, reduce operational costs and time-to-market with service delivery (see Cisco Crosswork Network Controller and Cisco Crosswork Network Services Orchestrator for more details).

We're also introducing new IOS XR Segment Routing innovations with Path Tracing, which can give customers hop-by-hop visibility into where packets are flowing to help detect and troubleshoot issues quickly and enable better customer outcomes on agility and cost reduction.

Another way Cisco is helping simplify networks is through our award-winning Cisco Routed Optical Networking architecture. By converging IP and optical layers, platforms such as the Cisco 8000, can support IP and private line services through coherent pluggable optics, advanced intelligence with segment routing, and multi-domain/multivendor automation with Crosswork Network Automation. We're striving to help our customers reduce costs while optimizing operations.

Use Cases

Given that traffic volumes are increasing, higher capacity is needed at the network intersection points, such as in the core. These core networks are in the IP backbone and metro regions, where we're seeing more traffic concentrating, as applications and services move closer to the user, user access speeds increase with fiber and 5G, and functionality such as peering, subscriber management and CDN get distributed locally.

To avoid traffic jams with network congestion, a scalable metro core is needed to transport all traffic types, particularly high-bandwidth latency-sensitive traffic. However, metro locations tend to be smaller with tighter space constraints, which is why space efficiency is critical. Scaling to 800G can help providers address space and traffic demands efficiently with metro applications.

At the same time, IP backbones that interconnect metro networks are important to scale and help reduce bottlenecks. According to Dell'Oro, upgrades with IP backbone networks represent the highest demand for 400G, since the Internet backbone includes both cloud and communications service provider networks that carry traffic with mobile, broadband, and cloud services.

Traffic volumes, which rose during the pandemic, haven't gone back to pre-pandemic levels as was expected, driven by remote/hybrid work and learning, which Dell'Oro believes is also driving the need for more network investment.  And as Sandvine points out, "the onslaught of video, compounded by a growing number of applications with greater demands for latency, bandwidth and throughput, is exerting extraordinary pressure on global networks".

As more people, applications, and devices get connected to global networks, more traffic continues to multiply in data centers, where we're also seeing higher capacity demands in spine/leaf environments, such as super-spine, in addition to Data Center Interconnect (DCI) and data center WAN/core networks. AI/ML workloads are different from traditional data center traffic because the processors are very high bandwidth devices that can overwhelm networks and impact job completion rates without sufficient spine capacity.  Dell'Oro also expects AI/ML workloads need 3x more bandwidth over typical workloads, with stringent requirements for lossless and low-latency networks.  As AI/ML clusters grow in system radix and capacity, they require denser spines that can efficiently scale to 28.8T with 72x400G ports in order to avoid chokepoints.

Internet For the Future at 800G Speeds

With our modular 800G systems, we can offer the flexibility to deploy dense Nx400G and Nx100G ports in various use cases and leverage our Flexible Consumption Model (FCM) that supports Pay-as-You-Grow (PAYG) licensing to help with budgeting goals over time.

We can help providers do more with Mass-scale Infrastructure for the Core, from enabling operational efficiencies to higher quality insightful experiences. Learn how with the award-winning Cisco 8000 Series.

1] Source: Based on Cisco internal study. Refer to Cisco 8000 Series Routers Data Sheet for device specifications.

[2] Source : Cisco internal lab testing, data sheet, power calculator, global emissions factor estimates, and Environmental Protection Agency  Greenhouse Gas Equivalencies Calculator