Skip to main content

Cloudscale routing for modern IP networks


The latest advances in IP routing silicon offer the power and opportunity to re-imagine IP networks — and deliver the reliable high performance that service providers and webscale operators need for success. In recent years, the need to boost network capacity and stay ahead of exponential traffic growth has dominated router design to the point that trade-offs between scaling capacity and features seem inevitable. Even important capabilities—including traffic management and security features—are being sacrificed for more capacity at lower cost. But now, with new intelligent routing silicon, compromises like this are no longer needed.

Instead, industry-leading capacity can be combined with real-time network intelligence on a single platform. So modern IP networks can deliver the secure high performance that’s required in the internet of things (IoT) and cloud era, but more cost-effectively, over a longer lifetime.

In other words, today’s leading IP routing silicon enables intelligence at scale—which is the key to enhancing QoE for highly diverse applications, while increasing network efficiency, service quality, operational security, and longevity.

Capabilities plus capacity: the modern IP network

Faced with an onslaught of new connected devices and fast-growing traffic volumes, the world’s leading service providers and webscale operators need routers with tremendous capacity and port density for their IP network architectures. However, scale alone is not sufficient, because using network capacity effectively requires a range of capabilities. And these capabilities need to be supported without compromising performance. That’s why IP routers must support distinct network roles and functions within service provider and webscale networks.

Service provider network functions include:

  • Broadband network gateway (BNG)
  • Provider edge
  • IP core
  • IP mobile anyhaul
  • Carrier Ethernet aggregation
  • Carrier Grade NAT (CGNAT)
  • WLAN gateway
  • Security gateway
  • Mobile packet core functions

Webscale network functions include:

  • Datacenter and WAN aggregation
  • Peering router
  • Datacenter gateway
  • Point of presence (PoP) edge router
  • Backbone router functions.

With the move to cloud-based architectures, video caching, mobile edge computing (MEC) functions, and datacenters (for private cloud, public cloud, and NFV-based applications) all need to be located closer to end users to help guarantee performance and QoE. This change drives the need for an intelligent and distributed high-performance network.

Faster broadband access speeds and the adoption of M2M/IoT also have an impact, because they increase security threats — requiring new approaches for mitigating attacks cost-effectively. The ability to apply security threat mitigation closer to end users or devices will be crucial for networks of the future.

Each of the network roles will therefore require high performance and scale for deep packet classification, IPv4 and IPv6 packet forwarding, routing policies and access control lists (ACLs), traffic management, and statistics with extensive packet counters. Additionally, multiple capabilities must often be supported concurrently without compromising performance.

Modern IP networks continue to push the limits of IP routing platforms and their ability to support massive scale, intelligence, and very high performance. Clearly IP routing silicon, which is the basic DNA for IP routers, needs to step to the challenge as well. But over the last few years, innovation has stagnated as silicon and memory development vendors have pulled out, and the industry’s focus has shifted toward creating systems for niche market segments such as DC switching and aggregation. Because these niche systems are mainly optimized for scale and limited capabilities, the shift leaves a huge void in the overall market.

The big challenge for IP networks has been—how to have it all, capacity plus capabilities. The solution has arrived through major advances in routing silicon.

The breakthrough value of intelligent silicon

As the networking industry has focused on scaling and hardware commoditization, silicon innovation has been in decline. Nokia has taken an alternative path, choosing, instead, to redefine scale in a way that eliminates design trade-offs.

This approach extends beyond offering sheer platform capacity or port density—and assumes that IP routing should support massive traffic volumes, but without compromising capabilities. Our engineers have made this combined power possible by creating “intelligent silicon” that enables massive router capacity, along with today’s most valuable capabilities. IP routing silicon can serve as a foundation for innovative capabilities based on insight-driven automation, and these capabilities are essential for delivering the highest QoE cost effectively.

Incorporating these network capabilities within IP routers is crucial in an era where the numbers and diversity of applications are expanding rapidly, with more apps and services having different cost and performance requirements. By eliminating the trade-offs between capacity and capabilities, modern IP routing can deliver better performance, improved customer satisfaction, and cost optimizations over time.

Like DNA, IP routing silicon is the key building block that ultimately defines a network’s potential—as well as the possibilities and limitations of the applications running on that network. If this foundation is weak, lacks key capabilities, or is not long-lasting, then the network, applications, end users, and network operators experience the consequences. And unlike software, silicon cannot easily be changed.

For instance, routers that emphasize capacity over capabilities often have port-level limitations, buffering constraints and performance degradation at scale. Or they may have anomalies in service consistency that end up requiring more complex network engineering and high-touch operations. In brief, a single-dimension product can’t really meet multiple needs. So what appears to be an economical choice, initially, often results in a more cumbersome and costly approach to delivering high volumes of network traffic.

In contrast, routers with intelligent silicon can be optimized in multiple dimensions. While providing enormous speed and capacity, they can also offer flexible new ways to promote reliable performance, efficient operations, enhanced security, easier management, and longer product life.

Why compromise when you can have it all?

To find out more about the value of intelligence at scale, please see the additional resources we have provided—including information on the new Nokia Deepfield solution.

Additional resources

Application note: Delivering intelligence at scale and extensibilityExecutive summary: IP networks reimaginedIP Networks Reimagined websiteNetwork Services PlatformDeepfield IP network analytics and DDoS protectionWhite paper: Insight-driven automated networkingAdaptive networks promote profits and security

Anthony Peres

About Anthony Peres

Anthony Peres has close to 25 years of experience in the IT/Telecommunications industry, with experience in IP/MPLS Network Design & Consulting, Product Management, Business Development, Solution Consulting and Product Marketing in various roles during his tenures at Wipro India, Newbridge Networks Canada and Alcatel-Lucent Canada. Anthony focuses on Alcatel-Lucent’s IP/MPLS products (Metro Aggregation, IP Edge and IP Core Routing) and service provider solutions in the residential, business and mobile segments. Anthony has completed his engineering studies in Mumbai, India and has an MBA from the University of Ottawa, Canada.

Jeff Jakab

About Jeff Jakab

A 19 year employee of Nokia, Jeff is presently a hardware PLM with a focus on technical marketing and strategic responsibilities within the 7750 SR, 7950 XRS, 7210 SAS and 7705 SAR product families. He presently works in the heart of Silicon Valley out of the Nokia office in Mountain View, California. Jeff has completed is undergraduate degree in Electrical Engineering with a Computer Engineering Minor from the University of Calgary, Canada, and also holds an MBA from the University of Ottawa, Canada.

Article tags