Bridging the Digital Divide with Subscriber Edge

Bridging the digital divide has been a longstanding top priority for countries globally. According to Broadband Research, in 2023 approximately five billion people (64% of the world’s population) were connected to the internet. That means roughly three billion people do not have the basic digital necessities such as access to data, sharing information, or communicating. In addition, they do not have the same access to educational, employment, and economic opportunities that could improve the quality of their lives through digital connections.

The World Bank has estimated that increasing the percentage of people with internet access to 75% would “boost the developing world’s collective GDP by $2 trillion and create 140 million new jobs.”

The good news is that the public and private sectors have been partnering to help close the digital divide, but as Broadband Research states: “Factors like increased affordability of devices, improved infrastructure and innovative services drive this growth.”

Role of subscriber edge

Accessing the internet requires a subscription to a broadband service from communications service providers (CSPs), using either cable, fiber, DSL, fixed wireless access (FWA), satellite or 4G/5G infrastructure and devices. Subscriber edge is the access point function for subscribers in a service provider network through which they connect to the broadband network for high-speed connectivity, such as accessing the internet.

Subscriber edge can be deployed with other services on the same platform by converging residential and enterprise services using multiservice nodes. Subscriber edge solutions involve managing subscriber sessions, and include functions like IP address management, policy and quality of service (QoS) enforcement, and secure access to the network through authentication and billing.

Shifting application and traffic demands

Traditional approaches for offering broadband services can be revenue-impacting and could affect the quality of experience (QoE) for a broadband user (see Figure 1). For example, with the advancement of applications and evolving transport protocols—such as Quick UDP Internet Connections (QUIC) and Transmission Control Protocol/ Transport Layer Security (TCPLS)—traffic patterns within broadband networks are experiencing a shift away from traditional transport protocols. These new protocols offer greater control to end-user applications, which reduces the dependency on the underlying broadband network and requires relatively simpler QoS models.

This shift is a pivotal opportunity for CSPs to simplify and modernize their complex traditional broadband networks to address higher bandwidth demands, growing user base, increasing video traffic and rising costs. As a result, there is a need to relook at subscriber edge with the overall subscriber services network design, and address important areas such as:

• Subscriber anchor point in the network
• Selection of subscriber edge devices and architecture
• Improve end-user experience and offer new services
• Address rising network costs

Figure 1. Traditional broadband centralized architecture

Source: cisco.com

Continue reading

Pave the Way for New Revenue with Transport Slicing Automation and Assurance

Excellence in service matters. Whether you are a mom-and-pop operation or a multibillion-dollar business spanning multiple industries, keeping customers happy and satisfied is essential. No one knows this better than telecommunication (telecom) providers—who experience close to 40% customer churn due to network quality issues, according to McKinsey.

For telecom providers, delivering an outstanding digital experience means smarter troubleshooting, better problem-solving, and a faster route to market for innovative and differentiated services. As if delighting customers and generating new revenue weren’t enough, there is also the added pressure of keeping operational costs low.

For operators of mass-scale networks, that can be a tall order. The good news is that innovative solutions like transport slicing and automated assurance are creating opportunities for service providers to build new revenue streams and differentiate services on quality of experience (QoE) with competitive service-level agreements (SLAs). In this blog post, we will explore how transport slicing and automated assurance can revolutionize the network landscape and transform service delivery, paving the way for financial growth.

Simplify and transform service delivery

Despite ongoing transformation efforts, telecom and high-performance enterprise networks are becoming increasingly complex and challenging to manage. Operations for multivendor networks can be even more complicated, with their various domains, multiple layers of the OSI stack, numerous cloud services, and commitment to end-to-end service delivery.

Complexity also impacts service performance visibility and how quickly you can find, troubleshoot, and fix issues before customer QoE is impacted. To understand the customer experience and differentiate services with competitive enterprise SLAs, you need real-time, KPI-level insights into network connections across domains and end-to-end service visibility. To confront these challenges, many service providers have been attempting to simplify network operations while reducing the cost of service delivery and assurance.

So, how do you streamline increasingly busy transport network operations? For one, advanced automation and orchestration tools can automate intent-based service provisioning, continuously monitor SLAs, and take corrective action to maintain service intent. Automated assurance, for example, can proactively monitor service performance, as well as predict and remediate issues before customers are impacted. This reduces manual work and allows for quicker reaction times to events that impact service.

With leading-edge platform and automation capabilities, you can deliver a wide range of use cases with a unified interface for visualization and control to manage network services effectively. This makes complex multidomain transport networks more accessible and easier to operate, which can help improve capital efficiency, enhance OpEx utilization, and accelerate new service launches.

The operational agility created by simplifying network operations allows you to adapt quickly to market changes and customer needs. You can assure services while securing new revenue streams and capitalizing on emerging service delivery opportunities.

Leverage transport slicing for the best outcomes

Simplifying network operations is only one part, because the one-size-fits-all approach to network infrastructure is no longer sufficient as your customers demand more personalized, flexible, and efficient services. This is where transport slicing can help, by offering a new level of customization and efficiency that directly impacts the service quality and service guarantees you can offer.

Network slicing is typically associated with delivering ultra-reliable 5G network services, but the advantages of transport slicing reach well beyond these areas. Network as a service (NaaS), for example, addresses enterprise customers’ need for more dynamic, personalized networks that are provisioned on demand, like cloud services. Using transport slicing, you can create customized virtual networks that provide customers with tailored services and control, while simplifying network management.

For telecom providers, transport slicing and automation will be critical to managing service level objectives (SLOs) of diverse, slice-based networks at scale. The transport layer plays a critical role in service delivery, and automation is essential to simplify operations and reduce manual workflows as slicing and enterprise services based on slicing become more complex.

Together, transport slicing and automation fundamentally change how network services are delivered and consumed. And while the slicing market is still in early stages of adoption, end-to-end slicing across domains offers a level of efficiency that translates into direct benefits through faster service deployment, enhanced performance, cost savings, and the ability to scale services quickly.

A complete, end-to-end automated slicing and assurance solution

Traditionally, the service monitoring and telemetry required to do any service-level assurance was an afterthought—built separately and not easily integrated into the service itself. Now, you can leverage the power of network slicing while ensuring each slice meets the stringent performance criteria your customers expect, as well as enable closed-loop automation based on end-user experiences at microsecond speeds. For example, tight integration between Cisco Crosswork Network Automation and Accedian’s performance monitoring solution provides a more complete, end-to-end automated slicing and assurance approach.

You can create and modify network slices based on real-time demand, and then leverage performance monitoring tools to not only assure the health and efficiency of these slices, but also provide empirical data to validate SLAs. You can use predictive analytics and real-time insights to identify and mitigate issues before they impact service quality, enabling increased network uptime and enhancing customer experience.

A proactive approach towards network management helps you use resources more efficiently, decrease complexity, and ensure customer satisfaction is prioritized—all while creating new opportunities for innovative revenue streams through highly differentiated and competitive service offerings. Ultimately, automated slicing and assurance drives greater operational excellence.

Fix problems before customers notice

Competition remains fierce in the telecom market, as service providers strive to meet B2B customers’ demand for high-quality services, fast service provisioning, and performance transparency. Speed is a differentiator, and customers notice immediately when service disrupts. Customers are willing to pay a premium for critical network performance, service insights, and enhanced SLAs that promise immediate resolution. To meet this demand, you need the ability to find and fix problems before customers notice.

Transport slicing and automated assurance are at the forefront of this challenge, enabling service providers to not only deliver services faster and more reliably, but to also have confidence those services will have the performance and QoE that customers expect.

The right network automation capabilities can drive simplified, end-to-end lifecycle operations, including service assurance that’s dynamic, intelligent, and automated. This paves the way for revenue-generating, premium services and delivering the outstanding experiences your customers expect.

Source: cisco.com

Continue reading

Transforming the Economics of Superfast Broadband with Cisco Routed PON

Today marks the launch of Cisco Routed PON, a truly disruptive solution that enables agile, differentiated broadband services through a software-defined broadband network. It’s part of our ongoing mission to transform the economics of networking for the benefit of communication service providers and communities worldwide. Routed PON drastically improves the cost of broadband deployment in rural, suburban, and urban areas, to help bring reliable, superfast connectivity to both residential and business customers.

In July 2016, the United Nations declared the internet a basic human right. Recognizing the importance of high-speed internet access in improving people’s lives and growing the digital economy, governments worldwide are investing heavily in broadband builds. The $42.45 billion Broadband Equity, Access and Deployment (BEAD) fund in the U.S. is just one example. Its goal is to ensure that every American can reap the benefits of high-speed internet access.

Communication service providers have welcomed initiatives like this because of the high cost of building new infrastructure and declining ARPU. Yet, bridging the digital divide and meeting both consumers’ and businesses’ growing bandwidth demands requires more than just public funding. It calls for a complete rethink of how broadband networks are built. That’s why we developed Cisco Routed PON—to help communication service providers and municipalities to deploy broadband networks in a better and simpler way.

Why can’t we just keep doing things the old way?

In today’s hyperconnected world—where hybrid work is the new normal, artificial intelligence (AI) innovation is accelerating, and new bandwidth-hungry applications continue to emerge—rolling out and managing profitable, high-performance broadband access networks is difficult and complex. And, it’s going to become even more difficult as bandwidth growth continues—from 10G, 25G and to 100G, and beyond.

The challenges are about connectivity and the services that broadband solutions enable. Our customers want to deliver services in an agile and cost-effective way, but they are increasingly constrained by traditional broadband architectures with large, dedicated optical line terminal (OLT) chassis that require dedicated space and power. Additionally, these chassis are separate from the access router, so they require separate layer management that can be costly. Traditional broadband architectures also offer less flexibility because they come as an integrated solution from a single vendor.

What sets Routed PON apart?

Unlike traditional chassis-based solutions, Cisco Routed PON enables communication service providers to put a small form factor PON pluggable in a router and converge FTTx access with their end-to-end network. It has three building blocks, all underpinned by a software-defined end-to-end architecture based on the IOS XR operating system.

1. Cisco Routed PON OLT Pluggable – A pluggable 10G OLT that replaces traditional stand-alone OLT chassis and connects the PON network to Layer 3 routing and services through a small form factor pluggable (SFP+) port on the router. The SFP is a cost optimized and power efficient way to deliver 10G symmetrical upstream and downstream data. Open and compliant with the OMCI standard, the OLT pluggable is compatible with any optical network terminal (ONT), helping customers avoid vendor lock-in.

2. Cisco Routed PON Controller – A stateless management controller that runs as a container on the router, configuring and monitoring end points in the PON network. It applies configurations to OLT and ONT devices and collects state information, statistics, alarms and logs from devices, and reports the information to higher layer applications.

3. Cisco Routed PON Manager – A WebUI application that acts as a graphical user interface for the PON network. The PON Manager facilitates device and service provisioning, and enables the management of users, databases, and alarms.

Flexibility, service differentiation, and investment protection

The capabilities of Cisco Routed PON lead to multiple positive business outcomes. The innovative architecture offers customers more flexibility because it’s interoperable with many ONTs. So, communication service providers can decide for themselves which ONT best meets their requirements and cost targets, upgrade to new features as needed, and not be tied to a single vendor’s roadmap.

Cisco Routed PON also makes their end-to-end architecture much simpler to manage, which in turn lowers OpEx. Instead of having separate systems and processes for PON, communication service providers can converge it with other access technologies on IP routers like active Ethernet – all unified by a common operating system, IOS XR, and automation.

At a time when reducing churn and growing revenue is critical, Cisco Routed PON helps customers stand out from competition and monetize their network investments in a smarter way. Thanks to its end-to-end architecture—with powerful IOS XR capabilities, such as segment routing and EVPN—it improves subscriber experience.

These capabilities also enable communication service providers to offer differentiated services for business and residential customers, such as ultra-low latency connectivity or additional security features. Crucially, Cisco Routed PON protects communication service providers’ investments as they build the Internet for the Future – ready for 10G, 25G, 50G, 100G, and beyond. When new higher-bandwidth Cisco pluggable OLTs become available, customers can simply plug them into their router on a port-by-port basis.

I’m proud of how Cisco keeps pushing the boundaries of routing and optical innovation to enable our customers to create more efficient and profitable network architectures. I see Cisco Routed PON as a further demonstration of how we are transforming and simplifying networking like we have done previously with Routed Optical Networking. I look forward to working with our customers as they leverage this new solution to accelerate the deployment of high-speed broadband in cities and rural communities around the world to bridge the digital divide.

Source: cisco.com

Continue reading

Showcasing Powerful Private 5G Use Cases at Cisco Live EMEA!

For those who joined us at Cisco Live! Amsterdam earlier this month, you might not have noticed that the venue featured a Private 5G Network established in partnership with NTT DATA.

Spanning two halls at RAI Amsterdam, or roughly 26,000 square meters, the seamless integration of this Private 5G network augmented the existing Wi-Fi network, pushing the boundaries of traditional connectivity, and creating a smart venue—a first for Cisco Live!

Built with the support of Intel, the Cisco Country Digital Acceleration team, and RAI Amsterdam—a conference and exhibition space that hosts millions of visitors annually—NTT DATA’s Private 5G network included four radios supporting mission critical and latency-sensitive applications. RAI also had over one hundred Wi-Fi access points supporting the user experience in the same location.

The entire ecosystem performed flawlessly. During busy hours with a full load on the network, Private 5G latency was a speedy 21.9 miliseconds, and Wi-Fi latency was 86 miliseconds. It was incredibly exciting to be part of the future of multi-access connectivity—wired and wireless, Wi-Fi, 4G and 5G, all brought together to enable a seamless digital experience.

The NTT DATA Private 5G-powered communication and streaming services were featured at Cisco Live! Amsterdam as part of the NTT DATA’s Smart Venue Solution, and included the following use cases:

• Mobile broadcasting – wireless video crews roamed the exhibition halls with low latency and high bandwidth, delivering a streamlined multi-camera environment.
• Visitor traffic management – Cisco’s Meraki cameras and NTT DATA’s Smart Management Platform tracked visitor movements and congestion, enabling operations and security teams to communicate real-time, data-driven crowd control decisions.
• Emergency Response Vehicle (ERV) – Pre-packaged, flexible FWA Private 5G connectivity was setup and used to mimic rural cellular/satellite backhaul.
• Premium Booth Connectivity – When the booth is already built and the floor is laid, network cable cannot be raised. P5G provided booth broadband for the exhibitor.
• NTT Coffee Booth – Cisco’s Meraki cameras and the NTT DATA’s Smart Management Platform monitored and managed queues and seating to optimize the on-site experience.
• Enhanced exhibitor experiences – Cisco’s Meraki cameras embedded throughout the venue and in booths captured anonymized data including the number of visitors and time spent in the booth to use for planning and to create better customer experiences.
• Out of Band management – The Private 5G network, backhaul connectivity, and network operations center were integrated to provide the Cisco Live! events team with faster coordination and emergency response capabilities.
• Venue Safety – Machine vision detected whether individuals were wearing Personal Protection Equipment (PPE) through a real-time alert system, helping to ensure safety throughout the convention center’s facilities.

Figure 1. NTT DATA Smart Venue Dashboard

Beyond the experience for event attendees, RAI benefited from the as-a-Service (aaS) model, which made it easy for them to “turn up” and support large amounts of data and real-time insights on the fly, seamlessly augmenting onsite experiences. Turning up 5G capabilities on an ad hoc basis is the ideal solution for conference centers that host large numbers of exhibitors and visitors.

Outfitting RAI with the ability to support advanced connectivity experiences was just the first step, our goal at Cisco is to provide our Service Provider customers with the seamless and flexible technology they need to create business outcomes that deliver on the bottom line.

According to Shahid Ahmed, Group Executive Vice President of New Ventures and Innovation at NTT DATA: “Private 5G and advanced analytics play a pivotal role in accelerating digitial transformation across industries and serve as a driving force to create smarter cities and venues. We are thrilled to partner with Cisco on this unique project. Private 5G excels in a complex environment like this one, and together with our Smart Management Platform will be the catalyst that accelerates the digital transformation journey for RAI and the City of Amsterdam.”

And the next steps at RAI? Cisco and NTT DATA plan to extend 5G coverage following Cisco Live to the venue’s vast 112,000 square meter footprint.

Source: cisco.com

Continue reading

Evolution to 5G-Advanced and Beyond: A Blueprint for Mobile Transport

The rapid rollout of 5G technology has marked a historic milestone in the evolution of mobile connectivity. According to research firm Omdia, 5G subscriptions surged from 1.4 billion in the middle of 2023 to a projected 8 billion by 2028, representing a compound annual growth rate (CAGR) of roughly 40%. Despite this impressive uptake, Omdia’s data also reveals that overall mobile revenue is growing at a modest rate of about 2%, and average revenue per user (ARPU) is experiencing a decline.

Wireless trends and opportunities

Communication service providers (CSPs) are responding by scaling their 5G networks to accommodate the soaring bandwidth demands, foster revenue growth, reduce total cost of ownership (TCO), and enhance network efficiency and agility.

The industry has seen significant investments from CSPs, with tens of billions of dollars spent on 5G spectrum and more on radio access network (RAN) infrastructure to support 5G. CSPs’ current focus is monetizing 5G for both consumer and enterprise services (see Figure 1).

Figure 1. Opportunities and Trends

On the consumer front, fixed wireless access (FWA) has emerged as a leading 5G application. For instance, in 2022, FWA accounted for 90% of net broadband additions in the U.S., surpassing traditional cable and DSL. However, this shift brings its own complexities, including the need for enhanced xHaul transport bandwidth, increased data center resources, and greater demand for spectrum resources.

For businesses, private wireless networks represent a crucial area of growth. These networks are particularly relevant in the manufacturing, transportation, logistics, energy, and mining sectors. The advent of 5G-Advanced technologies could help expand these opportunities further. Network slicing, introduced by the 3rd Generation Partnership Project (3GPP), will be pivotal in deploying private 5G networks and other differentiated services.

Partnerships are becoming increasingly important in network monetization strategies, especially with hyperscalers. Additionally, collaborations with satellite operators are gaining traction due to investment and dramatically reduced launch costs, enabling the deployment of low Earth orbit (LEO) constellations and satellite transition from proprietary silo towards integration with terrestrial and 5G networks. Driven by the need for comprehensive reachability and the development of standardized connectivity, as outlined in 3GPP Release 17, this collaboration allows mobile and fixed operators to expand coverage to remote locations and for satellite operators to tap into new customer bases.

Operators are also focusing on technical advancements to monetize their 5G networks effectively. This includes transitioning from non-standalone (NSA) to standalone (SA) mobile cores, which is essential for enabling advanced 5G capabilities. 5G SA cores are required to launch many capabilities supporting ultra-reliable low latency communications (URLLC), massive machine-type communications (mMTC), and network slicing.

Preparations are underway for 5G-Advanced (3GPP Release 18), with features like non-terrestrial networks (NTN), extended reality (XR), and advanced MIMO. The investment will be fundamental for advancing to 6G.

Another critical development is RAN decomposition and virtualization, which involves breaking down the RAN into individual components and running functions on commercial off-the-shelf hardware. Benefits include better utilization, greater scalability and flexibility, and cost reductions. Implementing decomposition and virtualization using O-RAN promises these benefits while breaking RAN vendor lock-in due to standardized, open interfaces.

Edge infrastructure investment is increasing to support new enterprise applications, integral to 5G SA and 5G-Advanced, by moving processing closer to end users, thereby reducing latency, and serving as a critical driver for cloud-native technology adoption. This approach requires flexible deployment of network functions either on-premises or in the cloud, leading to a decentralization of network traffic that was once concentrated. This evolving trend has become more pronounced with increasing traffic demands, blending network roles and boundaries, and creating a versatile network “edge” within the CSP’s framework.

Operational savings, including cost reduction and sustainability initiatives, are top priorities for CSPs to meet budgetary and carbon footprint goals.

Preparing your mobile transport for 5G Advanced and beyond

Mobile packet transport is critical in these initiatives and network transformation, leading to rapid changes in CSP transport networks. Traditionally, these networks relied on dedicated circuits and data communication appliances. However, modern transport is shifting toward a logical construct using any accessible hardware and connectivity services. Successful network architecture now hinges on the ability to seamlessly integrate a variety of appliances, circuits, and underlying networks into a unified, feature-rich transport network.

The Cisco converged, cloud-ready transport network architecture is a comprehensive solution designed to meet the evolving demands of 5G-Advanced and beyond. The architecture is particularly important for operators to navigate the complexities of 5G deployment, including the need for greater flexibility, scalability, and efficiency. Here’s a detailed look at its essential components:

• Converged infrastructure: Cisco’s approach involves a unified infrastructure seamlessly integrating various network services across wireline and wireless domains. This convergence is essential for supporting diverse customer types and services, from consumer-focused mobile broadband to enterprise-level solutions. The infrastructure is designed to handle all kinds of access technologies on a single network platform, including 4G, 5G, FWA, and the emerging direct satellite-to-device connectivity outlined in 3GPP’s NTN standards.
• Programmable transport and network slicing services: At the heart of Cisco’s architecture are advanced transport technologies like Border Gateway Protocol (BGP)-based VPNs and segment routing (SR), crucial for a unified, packet-switched 5G transport. These technologies enable a flexible services layer and an efficient underlay infrastructure. This layering provides essential network services like quality of service (QoS), fast route convergence, and traffic-engineered forwarding. Network slicing is also a key feature, allowing operators to offer customized, intent-based services to different user segments. This capability is vital for monetizing 5G by enabling diverse and innovative use cases.
• Cloud-ready infrastructure: Recognizing the shift toward cloud-native applications and services, Cisco’s architecture is designed to support a variety of cloud deployments, including public, private, and hybrid models. This flexibility ensures that the transport network can adapt to different cloud environments, whether workloads are on-premises or colocated. Virtual routers in the public cloud play a significant role here, providing required IP networking functions (including BGP-VPN, SR, and QoS).
• Secure and simplified operations model: Security and operational simplicity with service assurance are essential components in Cisco’s architecture. The network is designed for easy programmability and automation, which is essential for operational efficiency and cost reductions. This includes extensive telemetry and open APIs for easy integration with orchestration tools and controllers. Additionally, AI and machine learning technologies can potentially be used for real-time network visibility and actionable insights for optimizing user experience across both wireline and wireless networks.

The architecture is about current 5G capabilities and future readiness. Preparations for 5G-Advanced and the eventual transition to 6G are integral. The architecture’s design ensures operators can evolve their networks without major overhauls, thereby protecting their investment.

Cisco’s converged, cloud-ready transport network architecture offers a blend of technological innovation, operational efficiency, and flexibility, enabling operators to navigate the challenges of 5G deployment while preparing for the subsequent phases of network evolution.

Source: cisco.com

Continue reading

Award-Winning Centralized Platform Helps Unlock Value Through Simplicity

From work style to vehicle choice, hybrid has become the new norm. In fact, we are surrounded by use cases that need a hybrid approach to problem solving. And as we all know, networks are evolving. Today, networks need to be ready for new and growing applications such as artificial intelligence (AI), augmented and virtual reality (AR/VR), edge clouds, online gaming, connected cars, and video streaming. As a result, communication service providers (CSPs) are considering more options in redesigning networks.

For example, network operators need to cater to their customers by delivering services from anywhere between 1G to 100G speeds, while having the ability to aggregate into 400G networks. Operators need a platform that allows them to bridge this gap from 1G to 400G.

Platform design choices

Typically, there have been two types of form factors for routing platforms: fixed and distributed systems.

Fixed systems can contain a single forwarding chip and single route processor (RP) with fixed interfaces (see Figure 1). Fixed systems typically come in a “pizza box” form factor that is often used in network architectures that are more predictable and simpler, where using a system with fixed interfaces is suitable for anticipated network traffic patterns.

Figure 1. Fixed system

Distributed systems use a different architecture (see Figure 2), where the packet-forwarding decisions and actions take place on the network processor units (NPUs)/forwarding engines located on the individual line cards. Each card maintains a copy of the forwarding information base (FIB) that is distributed by the RP in the control plane. Large distributed systems have traditionally been designed to provide higher total system bandwidth and port densities, field-replaceable line cards, interface diversity, and redundancy.

These requirements have far exceeded what could be accomplished with a single NPU on a fixed system, which is why every line card has multiple NPUs participating in the forwarding decisions. This architecture helps deliver favorable customer outcomes with increased reliability and flexibility.

Figure 2. Distributed system

New hybrid choice with centralized architecture

With the evolution of the network and emergence of more localized and metro-driven traffic patterns, there is a need for network operators to deploy a solution that meets the needs of both fixed and distributed systems. Cisco 8000 Series Routers address this customer problem and market need by delivering a platform that is uniquely positioned to support the reliability and flexibility offered by distributed solutions, while also delivering value with the customer investments.

Instead of having to choose between a fixed or distributed system, customers can now also consider the new centralized system with Cisco 8600 Series Routers (see Figure 3), which blend the resource efficiency of fixed systems with the interface flexibility, upgradeability, and redundancy of distributed systems.

Figure 3. Centralized system

Similar to distributed systems, centralized systems have in-service, replaceable, redundant RPs with CPU and redundant switch cards (SCs) with NPUs to support both data plane and control plane redundancy. Cisco 8600 Series Routers have modular port adapters (MPAs) that can be replaced while in service and enable interface flexibility. Like fixed systems, the forwarding decisions on centralized platforms are handled centrally on the RP/SC instead of the line card.

With the unique centralized design of Cisco 8600 Series Routers, the life of a data packet is carefully managed such that when traffic ingresses on one of the MPA interfaces, the physical layer (PHY) on the ingress MPA sends the traffic to both SCs. The Silicon One ASIC on both SCs processes the packets, so in the event of a failure with the active SC, the other standby SC always has all the packets to support data plane redundancy. At a point in time, only the packets processed by the active SC are forwarded to the network, and packets processed by the standby SC are dropped.

Use cases

With currently over five billion global internet users, it is becoming increasingly impractical for capabilities such as peering to happen at only traditional, centralized internet exchanges. Distributed peering points are emerging across the network to help avoid unnecessarily backhauling traffic to centralized locations. However, metro locations such as colocation sites, data centers, and central offices can be space-constrained, and every additional rack unit (RU) of space is extremely costly.

Deploying right-sized platforms like Cisco 8600 Series Routers can address some of the operator resource challenges while achieving lower upfront costs, data plane and control plane redundancy, port diversity, and architectural simplicity using single-chip forwarding with less components to help lower TCO.

Additional use cases for the Cisco 8608 router include as a core label switch router (LSR), routed data center top-of-rack (ToR)/leaf, and aggregation for cloud and CSP networks. Cisco 8600 Series Routers are also part of the Cisco routed optical networking solution, with support for 400G DCO optics to improve network operational efficiency and simplicity.

Cisco innovations

Cisco Silicon One offers unmatched flexibility with a common silicon architecture, including software development kit (SDK) and P4 programmable forwarding code across multiple network roles (see Figure 4), while supporting fixed, distributed, and centralized systems (see Figure 5). With Cisco Silicon One used in Cisco 8600 Series Routers, we maintain the architectural simplicity and uniformity across the three architecture types. Having a unified architecture helps network operators simplify operations through consistency with upgrades, feature parity, training, testing/qualification, deployment, and troubleshooting.

Figure 4. Cisco Silicon One portfolio and network roles

Figure 5. Form factor types using Cisco Silicon One

Silicon One architecture achieves high performance and full routing capabilities without external memories. The clean-sheet internal architecture includes on-chip high-bandwidth memory (HBM) and supports multiple modes of operation by enabling a router to operate with a single forwarding chip, a line card network processor, and a switch fabric element. This flexibility enables consistent software experience in multiple roles and rapid silicon evolution.

Benefits of simplicity and uniformity across the three architecture types for network operators include:

• Consistent software experience across multiple network nodes.
• Simplified network operations through consistency with upgrades, qualification, deployment, and troubleshooting.
• Unified security and trust across the network.
• Programmable interfaces via consistent APIs.

In addition to the capabilities of the Silicon One chipset, Cisco 8600 Series Routers include significant innovations, such as the Cisco IOS XR network operating system (NOS) and the chassis design itself. For example, Cisco 8600 Series Routers enable all major components to be in-service field-replaceable, which helps reduce operational costs.

The single-forwarding chip design on Cisco 8600 Series Routers is well suited for smaller locations by offering simplicity through more bandwidth with fewer components, which helps streamline costs, power, and space (including with chassis depth of less than 600 mm) while also reducing latency.

The first platform in the Cisco 8600 Series Routers product line is the Cisco 8608 router, which includes these components:

• Chassis: The router has an eight-slot 7RU chassis at 580 mm depth, which hosts fans, power supplies, RPs, SCs, and MPAs.
• Route processor: The RP hosts the CPU complex and the I/O ports. RPs fit vertically in the chassis from the front panel. Up to two RPs are supported in the system and the RPs operate in active-standby mode for a redundant system.
• Switch card: SCs sit orthogonally in the back of the MPAs with connections to all MPAs. SCs directly host the NPUs, with up to two SCs in the system that work in active-standby mode to deliver data plane redundancy.
• Power supplies: The router has four power supplies that can provide redundant power to the system. The power options include pluggable 3.2 KW AC and pluggable 3.2 KW DC.
• Fans: There are eight fans in the system, with each fan individually removable or replaceable to provide N+1 fan redundancy to the system.
• Modular port adapters: With a high degree of flexibility, the Cisco 8608 router supports a diverse range of interfaces, including 4×400 GbE, 24×10/25/50 GbE, and a combination of 16×100 GbE or 12×100 GbE+1×400 GbE or 8×100 GbE+2×400 GbE.
• Network operating system: Cisco IOS XR is the common NOS across access, aggregation, edge, and core platforms, including Cisco 8600 Series Routers. IOS XR provides network intelligence, programmability, and trustworthy solutions to help deliver operational efficiency.
• Manageability: Cisco Crosswork Network Automation is a comprehensive software platform that helps plan, provision, manage, optimize, and assure multi-vendor/multi-domain networks, including Cisco 8600 Series Routers, to help reduce operational costs.

Customer benefits

The centralized architecture of Cisco 8600 Series Routers enables customers to take advantage of three main benefits (see Figure 6), including:

• Reliability: The unique hardware architecture provides industry-leading reliability with both control plane and data plane redundancy without loss of any front face plate.
• Flexibility: In-service upgradability and mix-and-match port support from 1G to 400G to help to efficiently meet both user and network traffic demands.
• Value: Customers can experience greater value with:
• Investment protection
• MPA backward compatibility
• Next-generation SC compatibility
• Optimized CapEx spending with right-sized platform to meet specific scale, space, power, and redundancy requirements
• Optimized OpEx spending with field-upgradeable and reusable components (similar to distributed systems) combined with using automated operations
• Sustainability that can help customers toward meeting their sustainability goals using a simplified centralized architecture.

Figure 6. Enabling customer outcomes

Meet evolving network priorities

Cisco is empowering customers with a hybrid architecture to meet their ever-changing network demands. Cisco 8600 Series Routers are a culmination of innovations in silicon, software, and hardware—all coming together to deliver a new breed of simple, reliable, flexible routers that give customers more choices and help maximize value.

Source: cisco.com

Continue reading

Safeguard Your Network in a Post-Quantum World

Security is critical when transmitting information over any untrusted medium, particularly with the internet. Cryptography is typically used to protect information over a public channel between two entities. However, there is an imminent threat to existing cryptography with the advent of quantum computers. According to the National Institute of Standards and Technology (NIST), “When quantum computers are a reality, our current public key cryptography won’t work anymore… So, we need to start designing now what those replacements will be.”

Quantum computing threat

A quantum computer works with qubits, which can exist in multiple states simultaneously, based on the quantum mechanical principle of superposition. Thus, a quantum computer could explore many possible permutations and combinations for a computational task, simultaneously and swiftly, transcending the limits of classical computing.

While a sufficiently large and commercially feasible quantum computer has yet to be built, there have been massive investments in quantum computing from many corporations, governments, and universities. Quantum computers will empower compelling innovations in areas such as AI/ML and financial and climate modeling. Quantum computers, however, will also give bad actors the ability to break current cryptography.

Public-key cryptography is ubiquitous in modern information security applications such as IPsec, MACsec, and digital signatures. The current public-key cryptography algorithms are based on mathematical problems, such as the factorization of large numbers, which are daunting for classical computers to solve. Shor’s algorithm provides a way for quantum computers to solve these mathematical problems much faster than classical computers. Once a sufficiently large quantum computer is built, existing public-key cryptography (such as RSA, Diffie-Hellman, ECC, and others) will no longer be secure, which will render most current uses of cryptography vulnerable to attacks.

Store now, break later

Why worry now? Most of the transport security protocols like IPsec and MACsec use public-key cryptography during the authentication/key establishment phase to derive the session key. This shared session key is then used for symmetric encryption and decryption of the actual traffic.

Bad actors can use the “harvest now, decrypt later” approach to capture encrypted data right now and decrypt it later, when a capable quantum computer materializes. It is an unacceptable risk to leave sensitive encrypted data susceptible to impending quantum threats. In particular, if there is a need to maintain forward secrecy of the communication beyond a decade, we must act now to make these transport security protocols quantum-safe.

The long-term solution is to adopt post-quantum cryptography (PQC) algorithms to replace the current algorithms that are susceptible to quantum computers. NIST has identified some candidate algorithms for standardization. Once the algorithms are finalized, they must be implemented by the vendors to start the migration. While actively working to provide PQC-based solutions, Cisco already has quantum-safe cryptography solutions that can be deployed now to safeguard the transport security protocols.

Cisco’s solution

Cisco has introduced the Cisco session key import protocol (SKIP), which enables a Cisco router to securely import a post-quantum pre-shared key (PPK) from an external key source such as a quantum key distribution (QKD) device or other source of key material.

Figure 1. External QKD as key source using Cisco SKIP

For deployments that can use an external hardware-based key source, SKIP can be used to derive the session keys on both the routers establishing the MACsec connection (see Figure 1).

With this solution, Cisco offers many benefits to customers, including:

• Secure, lightweight protocol that is part of the network operating system (NOS) and does not require customers to run any additional applications
• Support for “bring your own key” (BYOK) model, enabling customers to integrate their key sources with Cisco routers
• The channel between the router and key source used by SKIP is also quantum-safe, as it uses TLS 1.2 with DHE-PSK cipher suite
• Validated with several key-provider partners and end customers

Figure 2. Cisco SKS engine as the key source

In addition to SKIP, Cisco has introduced the session key device (SKS), which is a unique solution that enables routers to derive session keys without having to use an external key source.

Figure 3. Traditional session key distribution

The SKS engine is part of the Cisco IOS XR operating system (see Figure 2). Routers establishing a secure connection like MACsec will derive the session keys directly from their respective SKS engines. The engines are seeded with a one-time, out-of-band operation to make sure they derive the same session keys.

Unlike the traditional method (see Figure 3), where the session keys are exchanged on the wire, only the key identifiers are sent on the wire with quantum key distribution. So, any attacker tapping the links will not be able to derive the session keys, as having just the key identifier is not sufficient (see Figure 4).

Figure 4. Quantum session key distribution

Cisco is leading the way with comprehensive and innovative quantum-safe cryptography solutions that are ready to deploy today.

Source: cisco.com

Continue reading

Cisco Silicon One Breaks the 51.2 Tbps Barrier

In December 2019, we made a bold announcement about how we’d forever change the economics of the internet and drive innovation at speeds like no one had ever seen before. These were ambitious claims, and not surprisingly, many people took a wait-and-see attitude. Since then, we’ve continued to innovate at an increasingly fast pace, leading the industry with innovative solutions that meet our customers’ needs.

Today, just three and a half years after launching Cisco Silicon One™, we’re proud to announce our fourth-generation set of devices, the Cisco Silicon One G200 and Cisco Silicon One G202, which we are sampling to customers now. Typically, new generations are launched every 18 to 24 months, demonstrating a pace of innovation that’s two times faster than normal silicon development.

The Cisco Silicon One G200 offers the benefits of our unified architecture and focuses specifically on enhanced Ethernet-based artificial intelligence/machine learning (AI/ML) and web-scale spine deployments. The Cisco Silicon One G200 is a 5 nm, 51.2 Tbps, 512 x 112 Gbps serializer-deserializer (SerDes) device. It is a uniquely programmable, deterministic, low-latency device with advanced visibility and control, making it the ideal choice for web-scale networks.

The Cisco Silicon One G202 brings similar benefits to customers who still want to use the 50G SerDes for connecting optics to the switch. It is a 5 nm, 25.6 Tbps, 512 x 56 Gbps SerDes device with the same characteristics as the Cisco Silicon One G200 but with half the performance.

To achieve the vision of Cisco Silicon One, it was imperative for us to invest in key technologies. Seven years ago, Cisco began investing in our own high-speed SerDes development and realized immediately that as speeds increase, the industry must move to analog-to-digital (ADC)-based SerDes. SerDes acts as a fundamental building block of networking interconnect for high-performance compute and AI deployments. Today, we are pleased to announce our next-generation, ultra-high performance, and low-power 112 Gbps ADC SerDes capable of ultra-long reach channels supporting 4-meter direct-attach cables (DACs), traditional optics, linear drive optics (LDO), and co-packaged optics (CPO), while minimizing silicon die area and power.

Figure 1. Cisco Silicon One product family

The Cisco Silicon One G200 and G202 are uniquely positioned in the industry with advanced features to optimize real-world performance of AI/ML workloads—while simultaneously driving down the cost, power, and latency of the network with significant innovations.

The Cisco Silicon One G200 is the ideal solution for Ethernet-based AI/ML networks for several reasons:

~ With the industry’s highest radix switch, with 512 x 100GE Ethernet ports on one device, customers can build a 32K 400G GPUs AI/ML cluster with a 2-layer network requiring 50% less optics, 40% fewer switches, and 33% fewer networking layers—drastically reducing the environmental footprint of the AI/ML cluster. This saves up to 9 million kWh per year, which according to the U.S. Environmental Protection Agency is equivalent to more than 6,000 metric tons of carbon dioxide (CO2e) or burning 7.3 million pounds of coal per year.

~ Advanced congestion-aware load balancing techniques enable networks to avoid traditional congestion events.

~ Advanced packet-spraying techniques minimize creation of congestion hot spots in the network.

~ Advanced hardware-based link-failure recovery delivers optimal performance across massive web-scale networks, even in the presence of faults.

Figure 2. Benefits of large radix 51.2 Tbps switches

Cisco Silicon One Innovations

Here’s a closer look at some of our many Cisco Silicon One–related innovations:

Converged architecture

~ Cisco Silicon One provides one architecture that can be deployed across customer networks, from routing roles to web-scale front-end networks to web-scale back-end networks, dramatically reducing deployment timelines, while simultaneously minimizing ongoing operations costs by enabling a converged infrastructure.

~ Using a common software development kit (SDK) and standard Switch Abstraction Interface (SAI) layers, customers need only port the Cisco Silicon One environment to their network operating system (NOS) once and make use of that investment across diverse network roles.

~ Like all our devices, the Cisco Silicon One G200 has a large and fully unified packet-buffer optimizing burst-absorption and throughput in large web-scale networks. This minimizes head-of-line blocking by absorbing bursts instead of the generation of priority flow control.

Optimization across the entire value chain

~ The Cisco Silicon One G200 has up to two times higher radix than other solutions with 512 Ethernet MACs, enabling customers to significantly reduce the cost, power, and latency of network deployments by removing layers of their network.

~ With our own internally developed, next-generation, SerDes technology, the Cisco Silicon One G200 device is capable of driving 43 dB bump-to-bump channels that enable co-packaged optics (CPO), linear pluggable objects (LPO), and the use of 4-meter 26 AWG copper cables, which is well beyond IEEE standards for optimal in-rack connectivity.

~ The Silicon One G200 is over two times more power efficient with two times lower latency compared to our already optimized Cisco Silicon One G100 device.

~ The physical design and layout of the device is built with a system-first approach, allowing customers to run system fans slower, dramatically decreasing system power draw.

Innovative load balancing and fault detection

~ Support for non-correlated, weighted equal-cost multipath (WECMP) and equal-cost multipath (ECMP) load balancing capabilities with near-ideal characteristics help to avoid hash polarization, even across massive networks.

~ Congestion-aware load balancing for stateful ECMP, flow, and flowlet enables optimal network throughput with optimal flow-completion time and job-completion time (JCT).

~ Congestion-aware stateless packet spraying enables near ideal JCT by using all available network bandwidth, regardless of flow characteristics.

~ Support for hardware-based redistribution of packets based on link failures enables Cisco Silicon One G200 to optimize real-world throughput of massive scale networks.

Advanced packet processor

~ The Cisco Silicon One G200 uses the industry’s first fully custom, P4 programmable parallel packet processor capable of launching more than 435 billion lookups per second. It supports advanced features like SRv6 Micro-SID (uSID) at full rate and is extendable with full run-to-completion processing for even more complex flows. This unique packet processing architecture enables flexibility with deterministic low latency and power.

Deep visibility and analytics

~ Programmable processors enable support for standard and emerging web-scale in-band telemetry standards providing industry-leading network visibility.

~ Embedded hardware analyzers detect microbursts with pre- and post-event logging of temporal flow information, giving network operators the ability to analyze network events after the fact with hardware time visibility.

A new generation of network capabilities

Gone are the days when the industry operated in silos. With its one unified architecture, Cisco Silicon One erases the hard dividing lines that have defined our industry for too long. Customers no longer need to worry about architectural differences rooted in past imagination and technology limitations. Today, customers can deploy Cisco Silicon One in a multitude of ways across their networks.

With the Cisco Silicon One G200 and G202 devices, we extend the reach of Cisco Silicon One with optimized high-bandwidth devices purpose-built for spine and AI/ML deployments. Customers can save money by deploying fewer and more efficient devices, enjoy new deployment topologies with ultra-long-reach SerDes, improve their AI/ML job performance with innovative load balancing and fault discovery techniques, and improve network debuggability with advanced telemetry and hardware analyzers.

If you’ve been watching since we first announced Cisco Silicon One in December 2019, it is easy to see that this is just the beginning. We’re looking forward to continuing to accelerate the value addition for our customers.

Stay tuned for more exciting Cisco Silicon One developments.

Source: cisco.com

Continue reading

Building AI/ML Networks with Cisco Silicon One

It’s evident from the amount of news coverage, articles, blogs, and water cooler stories that artificial intelligence (AI) and machine learning (ML) are changing our society in fundamental ways—and that the industry is evolving quickly to try to keep up with the explosive growth.

Unfortunately, the network that we’ve used in the past for high-performance computing (HPC) cannot scale to meet the demands of AI/ML. As an industry, we must evolve our thinking and build a scalable and sustainable network for AI/ML.

Today, the industry is fragmented between AI/ML networks built around four unique architectures: InfiniBand, Ethernet, telemetry assisted Ethernet, and fully scheduled fabrics.

Each technology has its pros and cons, and various tier 1 web scalers view the trade-offs differently. This is why we see the industry moving in many directions simultaneously to meet the rapid large-scale buildouts occurring now.

This reality is at the heart of the value proposition of Cisco Silicon One.

Customers can deploy Cisco Silicon One to power their AI/ML networks and configure the network to use standard Ethernet, telemetry assisted Ethernet, or fully scheduled fabrics. As workloads evolve, they can continue to evolve their thinking with Cisco Silicon One’s programmable architecture.

Figure 1. Flexibility of Cisco Silicon One

All other silicon architectures on the market lock organizations into a narrow deployment model, forcing customers to make early buying time decisions and limiting their flexibility to evolve. Cisco Silicon One, however, gives customers the flexibility to program their network into various operational modes and provides best-of-breed characteristics in each mode. Because Cisco Silicon One can enable multiple architectures, customers can focus on the reality of the data and then make data-driven decisions according to their own criteria.

Figure 2. AI/ML network solution space

To help understand the relative merits of each of these technologies, it’s important to understand the fundamentals of AI/ML. Like many buzzwords, AI/ML is an oversimplification of many unique technologies, use cases, traffic patterns, and requirements. To simplify the discussion, we’ll focus on two aspects: training clusters and inference clusters.

Training clusters are designed to create a model using known data. These clusters train the model. This is an incredibly complex iterative algorithm that is run across a massive number of GPUs and can run for many months to generate a new model.

Inference clusters, meanwhile, take a trained model to analyze unknown data and infer the answer. Simply put, these clusters infer what the unknown data is with an already trained model. Inference clusters are much smaller computational models. When we interact with OpenAI’s ChatGPT, or Google Bard, we are interacting with the inference models. These models are a result of a very significant training of the model with billions or even trillions of parameters over a long period of time.

In this blog, we’ll focus on training clusters and analyze how the performance of Ethernet, telemetry assisted Ethernet, and fully scheduled fabrics behave.

AI/ML training networks are built as self-contained, massive back-end networks and have significantly different traffic patterns than traditional front-end networks. These back-end networks are used to carry specialized traffic between specialized endpoints. In the past, they were used for storage interconnect, however, with the advent of remote direct memory access (RDMA) and RDMA over Converged Ethernet (RoCE), a significant portion of storage networks are now built over generic Ethernet.

Today, these back-end networks are being used for HPC and massive AI/ML training clusters. As we saw with storage, we are witnessing a migration away from legacy protocols.

The AI/ML training clusters have unique traffic patterns compared to traditional front-end networks. The GPUs can fully saturate high-bandwidth links as they send the results of their computations to their peers in a data transfer known as the all-to-all collective. At the end of this transfer, a barrier operation ensures that all GPUs are up to date. This creates a synchronization event in the network that causes GPUs to be idled, waiting for the slowest path through the network to complete. The job completion time (JCT) measures the performance of the network to ensure all paths are performing well.

Figure 3. AI/ML computational and notification process

 

This traffic is non-blocking and results in synchronous, high-bandwidth, long-lived flows. It is vastly different from the data patterns in the front-end network, which are primarily built out of many asynchronous, small-bandwidth, and short-lived flows, with some larger asynchronous long-lived flows for storage. These differences along with the importance of the JCT mean network performance is critical.

To analyze how these networks perform, we created a model of a small training cluster with 256 GPUs, eight top of rack (TOR) switches, and four spine switches. We then used an all-to-all collective to transfer a 64 MB collective size and vary the number of simultaneous jobs running on the network, as well as the amount of network in the speedup.

The results of the study are dramatic.

Unlike HPC, which was designed for a single job, large AI/ML training clusters are designed to run multiple simultaneous jobs, similarly to what happens in web scale data centers today. As the number of jobs increases, the effects of the load balancing scheme used in the network become more apparent. With 16 jobs running across the 256 GPUs, a fully scheduled fabric results in a 1.9x quicker JCT.

Figure 4. Job completion time for Ethernet versus fully scheduled fabric

Studying the data another way, if we monitor the amount of priority flow control (PFC) sent from the network to the GPU, we see that 5% of the GPUs slow down the remaining 95% of the GPUs. In comparison, a fully scheduled fabric provides fully non-blocking performance, and the network never pauses the GPU.

Figure 5. Network to GPU flow control for Ethernet versus fully scheduled fabric with 1.33x speedup

 

This means that for the same network, you can connect twice as many GPUs for the same size network with fully scheduled fabric. The goal of telemetry assisted Ethernet is to improve the performance of standard Ethernet by signaling congestion and improving load balancing decisions.

As I mentioned earlier, the relative merits of various technologies vary by each customer and are likely not constant over time. I believe Ethernet, or telemetry assisted Ethernet, although lower performance than fully scheduled fabrics, are an incredibly valuable technology and will be deployed widely in AI/ML networks.

So why would customers choose one technology over the other?

Customers who want to enjoy the heavy investment, open standards, and favorable cost-bandwidth dynamics of Ethernet should deploy Ethernet for AI/ML networks. They can improve the performance by investing in telemetry and minimizing network load through careful placement of AI jobs on the infrastructure.

Customers who want to enjoy the full non-blocking performance of an ingress virtual output queue (VOQ), fully scheduled, spray and re-order fabric, resulting in an impressive 1.9x better job completion time, should deploy fully scheduled fabrics for AI/ML networks. Fully scheduled fabrics are also great for customers who want to save cost and power by removing network elements, yet still achieve the same performance as Ethernet, with 2x more compute for the same network.

Cisco Silicon One is uniquely positioned to provide a solution for either of these customers with a converged architecture and industry-leading performance.

Figure 6. Evolve your network with Cisco Silicon One

Source: cisco.com

Continue reading