Cisco Network Convergence System: Building the Foundation for the Internet of Everything

1
Executive Summary
A third wave of Internet expansion, the Internet of Everything (IoE), is
expected to come through the confluence of people, processes, data, and
things. Cisco estimates the Value at Stake for the IoE is $14.4 trillion over
the next 10 years. IoE requires a context-intelligent architecture in addition
to massive scale. For IoE, service providers must deliver increasing security,
reliability, and availability with networks that also provide greater service
velocity and flexibility. In addition, they require network autonomics,
process automation and the tools to meet the management challenges of a
much more complex network.
Cisco Network Convergence System (NCS) is a family of integrated packet
routing and transport systems designed to help service providers capture
their share of the IoE Value at Stake. NCS is built on major innovations in
silicon, optics and software and provides the building blocks of a multilayer
converged network that intelligently manages and scales functions across its
architecture.
ACG Research analyzed the business case for NCS and found it achieves
massive scale via multichassis system architecture, the density and
performance of its new chip set, and the extension of the control plane to
virtual machines (VM) internally and externally. Fully virtualized software
improves service velocity and asset utilization by creating a cloud model
inside the platforms. Virtualization also supports orders of magnitude
improvement in system availability and security through the isolation and
independence of software operations. Optical innovations lower
multichassis interconnect costs and optimize wavelength density and cost.
Cisco Network Convergence System: Building
the Foundation for the Internet of Everything
KEY FINDINGS
NCS provides the architecture
for building multilayer
converged networks that
intelligently split and scale
functions to enable IoE:
Providing a system with up to
66% lower TCO than
competing solutions.
Consuming up to 80% less
power than competitive offers.
Accelerating service velocity
through fully virtualized
software.
Increasing system availability
beyond five 9s by isolating
software modules through
virtualization.
Opening the system to cloud-
based innovations and cost
levels by extending the NCS
control plane to VMs, both
internal and external.

2
Introduction
A third wave of Internet expansion, the Internet of Everything (IoE), is expected through the confluence
of people, processes, data, and things (Figure 1).
Figure 1 – Third Wave of Internet Development1
IoE opportunities are being driven by dramatic increases in processing power, storage, and bandwidth at
lower cost; the rapid growth of cloud, social media, and mobile Internet; the capacity to analyze large
quantities of data, transactions, and events and turn it into actionable intelligence (Big Data); and an
improved ability to use hardware and software more powerfully and flexibly with capabilities such as
virtualization, software-defined networking, and network function virtualization (NFV). IoE creates value
by leveraging large numbers of interconnected information flows. Cisco estimates the global IoE Value at
Stake2
over the next 10 years will be $14.4 trillion. Service providers can gain a share in IoE Value at
Stake by enabling the IoE.
IoE Driven Requirements for Service Providers’ Networks
IoE requires an evolved, programmable, context-intelligent architecture in addition to massive and
adaptive scale. Such a network must dynamically manage the complex interactions among traffic,
business rules, and transactions. Network requirements created or amplified by IoE include:
Interconnect trillions of things: About 10 billion high-function devices are connected to the
Internet today. The addition of vast numbers of things (such as individual retail and commercial
items) to the people on the Internet, the data they create and access, and the processes that
manage, analyze, and respond to all the data and communications requires several orders of
magnitude increase in connectivity.
1
See “Internet of Everything: Global Private Sector Economic Analysis,” Cisco, January 2013.
2
Cisco defines IoE Value at Stake as the total of private sector cost reductions and new revenue creation
attributable to IoE initiatives. See “Internet of Everything: Global Private Sector Economic Analysis,” Cisco, January
2013.

3
Secure private data: IoE adoption requires that information owners trust the Internet to protect
the owners’ information. This trust level requires much safer security measures than those
existing on the Internet today.
Support exponential traffic growth: Video, mobile broadband, cloud computing, and IoE will
drive global IP traffic to 1.4 zettabytes by the end of 2017 according to the Cisco Visual
Networking Index. Traffic is expected to maintain a 23 percent annual growth rate through
2017.
Increase network reliability and availability: IoE depends upon always-on networks. Today, the
five 9s availability criterion is not adequate for such critical tasks as process control, public
safety systems or medical monitoring.
Increase service velocity and flexibility: Service innovation must be fast and flexible. Service
introduction must be in days or minutes not months. Service instantiation must be low cost with
fast and flexible service chaining.
Meet increasing network management challenges: Managing huge numbers of devices, services
and protocols; complex security requirements; and massive scalability requires new approaches.
Automated network interactions will create self-managing and adaptive networks to enable
growth far beyond the size of the Internet today. Such autonomic networks are analogous to the
nervous system that regulates bodily functions and organs without conscious control.
The NCS Family
The NCS family of integrated packet and transport systems positions service providers to capture a share
of the IoE Value at Stake by meeting the new requirements for service providers’ networks (See Figure
2).
Figure 2 – NCS Product Family

4
NCS enables multilayer converged networks to split and scale functions intelligently across its system
architecture. The NCS family consists of three products:
NCS 6000: Integrated optical routing system
NCS 4000: Integrated packet optical transport system
NCS 2000: Agile DWDM system
NCS is built on major innovations in silicon, optics and software. A new level of silicon integration
replaces 14 chips with two. It features common building blocks used in a variety of combinations for
different applications, and all have programmable microcode. This enables rapid innovation and product
delivery.
Optical innovations feature 100 Gbps coherent detection technology3
and InfiniBand technology to
lower multichassis interconnect costs. Silicon photonics is used to optimize wavelength density and
power. The DWDM system is colorless, contention-less and omnidirectional. This simplifies network
management, accelerates network reconfiguration, and improves asset utilization.
NCS creates a paradigm shift in network systems software architecture. For the first time full software
virtualization extends down to the line card (See Figure 3).
Figure 3 – Fully Virtualized Software
The virtualization feature of NCS allows more flexibility than solutions of previous generations.
Hypervisors running VMs support a cloud model inside NCS. Figure 3 shows control plane, data plane
and network services (IP mobile core, for example) deployed on identical system software
infrastructure. Independent kernels and resources provide an isolated, scalable, and distributed owner-
tenant model within the system and extend to any external VM platform.
The extension of the control plane to VM environments by NCS, including Cisco UCS, allows almost
infinite flexibility and scalability of the routing control plane. This means new protocols, functions, and
3
See ACG Research’s “Why Service Providers Should Consider IPoDWDM for 100G and Beyond” for an analysis of
the economic benefits of Cisco’s coherent detection technology.

5
services can be deployed quickly, increasing service velocity and thus directly accelerating top-line
revenue.
The ability to run multiple concurrent versions of software reduces security risks, increases service
availability and service velocity. Virtualization also increases asset utilization, enables on-demand
delivery and further enhances in-service software upgrades. Extension of the control plane to VM
environments delivers the granular scale of customized services and low cost of data center-based
systems. This allows the flexibility to add new protocols and services without needing to upgrade the
hardware or adversely affecting performance. It provides a responsive low-cost platform for NFV,
opening access to a large number of innovative applications from an ecosystem of third-party
developers.
NCS Meets IoE Driven Requirements for Service Providers
NCS delivers the scale to support massive and unpredictable traffic growth. It supports Terabit per
second line cards with the highest 10 GE and 100 GE port densities in the industry, all providing line-
speed data rates. The extensibility of its control plane to VM environments gives NCS the elasticity to
respond to the dynamic demands of the IoE. Adding that to its network processing unit performance
and programmability yields a system architecture built to support the transaction rates of trillions of
connections and to evolve to meet future requirements.
Massive Scale via Multichassis Network System Architecture
The multichassis system architecture of NCS supports enormous traffic growth without additional
systems and interconnects between them. The architecture also eliminates the time and cost required
to reassign router ports when a new router is added to a cluster.
It is economically inefficient to accommodate traffic growth by adding single chassis routers to a router
cluster. Some router port capacity is consumed to interconnect each existing router to the additional
router. As more routers are added:
Interconnect ports displace more revenue-generating service ports.
Service ports need to be moved from one router to another, disrupting service and increasing
operational complexity and cost.
To meet traffic requirements, interconnect ports use between one-third and one-half of the
port capacity of the routers.
The interconnect architecture is a blocking-switching architecture and in some traffic conditions
may not have enough bandwidth to serve demand.
NCS resolves these problems through use of multiple line card chassis to accommodate line cards (and
first and third stage switch fabric cards) and a fabric card chassis that hosts the second stage of switch
fabric cards. The NCS architecture delivers:
No service disruption, no loss of service ports, and no rewiring when adding line card chassis

6
Nonblocking fabric design using high-speed optical interconnect with cost-reduced, pluggable
optics
System scale that back-to-back or clustering solutions cannot match
TCO and Power Comparisons with Leading Competitors
A case study is used to compare the total cost of ownership (TCO) and power consumption of NCS 6000
with Vendor A’s single router solution and Vendor B’s multichassis router solution. Both vendors’
solutions use their latest core routers. In the case study 2,000 10 GE ports and 500 100 GE ports are
required in the initial study year, and total port requirements grow at 10 percent CAGR over the next
five years. Figure 4 shows the TCO comparison; Figure 5 shows the power comparison.
Cisco NCS has 18 percent lower TCO than Vendor A and 45 percent lower TCO than Vendor B over six
years. Over the same study period Cisco uses 42 percent less power than Vendor A and 59 percent less
power than Vendor B. These savings are primarily attributable to the greater 100 GE and 10 GE port
densities of the NCS 6000 and to the highly integrated and power-efficient design of the NCS chip set.
Figure 4 – Core Router Six-Year Cumulative TCO Comparison
Figure 5 – Core Router Annual Power Consumption

7
TCO Benefits of Scalable Control Plane
NCS software virtualization combined with an extendable Ethernet control plane allows NCS networks
almost infinite control plane scalability. This is very important because some network services require
both control plane scalability and data plane scalability. A good example is the widely deployed MPLS
VPN service. MPLS VPNs stress the control plane with multiple routing tables and eBGP sessions. Typical
high-end routers cannot support more than 5,000 MPLS sessions per router. This is a limit imposed by
the control plane. Therefore, while large routers can scale to terabits per second on the data plane, the
control plane service limits can be exceeded long before the data plane limits are exceeded.
As an example, Figure 6 compares the NCS 6000 with core routers from Vendor A and B for a MPLS VPN
use case.
Figure 6 – TCO Comparison for MPLS VPN Control Plane Use Case
In this example each VPN customer’s access point is assumed to generate 50 Mbps of traffic on average.
The control plane limits for this service are exceeded long before the data plane limits are exceeded.
Using the NCS 6000, servers are added to the control plane to continue to scale the MPLS VPN service
without adding additional NCS chassis. Vendors A and B, however, require new chassis once the control
plane limits are exceeded even though there are empty slots and unused capacity in the data plane. In
this scenario the NCS 6000 solution is 66 percent more cost effective than Vendor A and 54 percent
more cost effective than Vendor B. The NCS 6000 uses 80 percent less power than Vendor A and 65
percent less power than Vendor B.
IP + Optical Convergence TCO Benefits
Cisco nLight, a multilayer routing and optimization architecture focused on IP and DWDM integration
and a Cisco ONE element, embodies the autonomics concept. nLight enables information flows between
the optical and routing layers via control plane interworking, delivering an end-to-end protection and
restoration solution. Compared to 1+1 optical protection nLight delivers savings of4
:
26 percent fewer router interfaces
64 percent fewer optical transponder interfaces
4
“The Economics of Cisco’s nLight Multilayer Control Plane Architecture,” ACG Research, 2013, provides the full
analysis of the savings produced by nLight.

8
Up to 50 percent savings in TCO for capacity planning
Optical Port Density Benefits
In addition to the TCO benefits of IP + optical convergence, there are additional operation (OpEx)
benefits associated with the high DWDM port density in the NCS 4000. A single NCS 4000 shelf can
support 32 100GE IPoDWDM interfaces compared with 16 or fewer 100GE interfaces provided by
competing solutions. Additionally, power consumption per Gbps in the NCS 4000 is half that of the
competition. This directly leads to 50 percent reductions in power, cooling, and floor space OpEx.
Reliability, Availability, Flexibility, Security, and Network Management Advantages
Full software virtualization allows software components to be isolated from each other. This protects
each software component from failures or security events elsewhere in the system and provides more
overall network reliability. Benefits are faster service velocity, including faster deployment of new
software features and functions; zero packet loss and zero topology loss during software upgrades; and
improved security.
Faster Service Velocity
Virtualization supports the rapid development and introduction of new services and software features
by isolating the new software in production and by abstracting functions from the underlying physical
infrastructure. An ACG Research study5
found software virtualization significantly reduces time for
testing and deploying new services and software. Substantial savings can be achieved by virtualizing only
the testing process. For example, an ACG Research study found 30 percent time savings and 90 percent
TCO savings by virtualizing a network test lab.
Zero Packet Loss, Zero Topology Loss and Network Availability
As IoE becomes more integral to business operations the need for more stringent service level
agreements (SLA) will grow. Their value will increase because of the higher impact of outages, and
remedies will escalate along with potential impact. Maintenance windows included in SLAs are typically
two hours per month. This alone brings network availability down to 0.997 (24 hours per year): an
availability level inadequate to meet the requirements of many IoE applications. Process control and
monitoring applications require availability objectives with annual outage time in the second or
subsecond range. Fault detection and isolation (protective relaying) used on the power grid is one such
example. Grid performance requires faults to be isolated within three electric current cycles (50 ms) in
order to minimize two catastrophic outcomes. First, faults across 500,000 volt transmission lines cause
catastrophic damage if they are of longer duration. Second, faults of longer duration can cascade across
the power grid, bringing down service over entire regions of a country6
.
5
See ACG Research’s “Economic Benefits of Cisco CloudVerse,” an analysis of how cloud techniques such as
virtualization, abstraction, orchestration, and automation increase service velocity. These techniques are enabled
by NCS’s fully virtualized architecture.
6
The 1999 southern Brazil blackout was caused by a lightning strike at a substation and cascading effect that
affected 75 to 97 million people. The Northeast U.S. blackout of 2003 affected 55 million people in the U.S. and
Canada and contributed to the deaths of 11 people. Water supply, transportation, communications and industry
also were impaired due to power failure.

9
The high availability architecture of NCS will be an essential requirement in the era of IoE. Virtualization
and isolation of software components eliminate packet and topology losses during software upgrades.
This efficiency increases network availability, eliminates downtime costs, and eliminates the need to
include maintenance window exceptions in SLAs.
Another way to assess the value of network availability is to observe the credits (remedies) service
providers apply to enterprise service subscribers’ accounts for violation of SLA network availability
provisions. Table 1 provides some examples.
SLA – Business Ethernet
Service
Remedy
Availability 99.99% 1% of MRC for each 1% below SLA
Latency 110 ms 1% of MRC for every 4 ms above SLA
Packet Delivery 99.9% 1% of MRC for each 1% below SLA
MRC = Monthly Recurring Charge
Table 1 – Remedies for Violation of SLA Availability Provisions7
Service providers can build networks to meet the SLA demands of IoE use cases with the innovative NCS
architecture as their foundation.
Improved Network Security
The isolation of software modules and operating systems (OS) of NCS improves network security. If, for
example, a hacker breaks into the data plane OS he does not have access to the control plane OS or the
administrative OS, because these are separate virtual operating systems. Additionally, the software
architecture of NCS allows many types of security software to run on the data, control, and
administrative planes of NCS. Examples include anti-DDOS, firewall, and intrusion protection system
services.
A simple probability model shows how software isolation decreases the chance that the entire network
will be compromised by a security attack. A router might host three software modules and employ three
operating systems. Assume there is a 0.01 probability that each of these modules and OSs can be
compromised by a security attack. Under monolithic software design there is a 0.06 chance the entire
router will be compromised. If each software module and OS is isolated, however, then there is only a
10-12
chance the entire router will be compromised8
.
The value of network security to service providers can be estimated by observing that recurring fees for
managed security services make up 20 percent of total charges paid by enterprises.9
This share of total
revenue is likely to increase as security needs increase with the huge growth in connections and with
more critical interdependent services riding the infrastructure.
7
Source ACG Research survey 2013
8
Under monolithic software probability that the entire router is compromised = 1 - (1 – 0.01)
6
= 0.06. Under
isolated software probability that entire router is compromised = 0.01
6
= 10
-12
.
9
Source ACG Research survey 2013

10
Benefits of NCS and Cisco ONE
Cisco ONE provides a comprehensive framework of modular technologies and network programmability
that allows networks to interact more meaningfully with applications to deliver on-demand response to
their requests, accelerated service deployment, and improved customer experience. The NCS system is
an integral part of the programmable network and its closed-loop communications path of analytics,
applications and services orchestration. Through the Cisco Developer Network, Cisco is enabling a ONE
ecosystem by helping developers create, test, and promote applications for their large customer base;
more than 75 developers, including SAP and Citrix, are actively developing with Cisco APIs.
Autonomics
The aim of autonomic networking is to create self-managing networks with automated reprogramming
and configuration of services in real time to overcome the rapidly growing complexity of the Internet
(especially IoE) and to enable its future growth far beyond the size it is today. The virtualized
architecture of NCS and the Cisco ONE ecosystem meet the challenges of creating autonomic networks.
IoE Applications
The Cisco ONE ecosystem provides a vehicle enabling service providers to enhance industry-specific IoE
initiatives. Some of the most attractive use cases include:
Smart factories: Adding connectivity to manufacturing processes and applications increases
factory productivity, reduces inventories with real-time inventory supplies, and cuts average
production and supply-chain costs.
Connected marketing and advertising: Broad IT and social applications for marketing and
advertising transform the way companies engage with customers, analyze their behavior, and
optimize the impact of their interactions.
Smart grid: An effective smart grid uses network connections—from production to customers—
to better understand users’ behaviors and improve the reliability, economics, and sustainability
of the production and distribution of electricity.
Real-time Analytics
The IoE concept is tightly coupled to Big Data—data sets too big for traditional database tools or data
processing methods. Network analytics increases the value of data when it can be used proactively in
real time to drive actionable events for modern business processes (for example, policy management).
Opportunities include marketing and sales for revenue-generating use cases, network management for
root-cause analysis and fault prediction, and security for fraud detection.
Cisco ONE optimizes the use of network resources by integrating network analytics into the production
network and enabling the network to adapt dynamically to application and user needs.
Conclusion
The IoE is creating tremendous opportunity for service providers to share in a $14.4 trillion Value at
Stake over the next 10 years. To do so service providers’ networks must meet its many new and
changing requirements:

11
Interconnect trillions of things
Secure private data
Support exponential traffic growth
Increase network reliability and availability
Increase service velocity and flexibility
Meet increasing network management challenges
Support autonomics and process automation
Cisco NCS is a family of integrated packet and transport systems that help service providers capture a
share of the IoE Value at Stake. NCS is built on major innovations in silicon, optics and software. Its new
chip set provides the foundation to handle the traffic and transactions trillions of connections will bring.
Optical innovations lower multichassis interconnect costs and optimize wavelength density and cost.
The fully virtualized software system integrates cloud benefits into the solution, enabling elastic
extension of the control plane, improving service velocity and asset utilization, and dramatically
increasing system availability and security. The multichassis system architecture of NCS builds multilayer
converged networks that intelligently manage, adapt, and massively scale functions to unlock IoE value
for service providers.
ACG Research is an analyst and consulting company that focuses in the networking and telecom space. Our best-in-
class subject matter analysts have a combined 120+ years of experience and expertise in telecom segments that
address both technology and business issues. We offer comprehensive, high-quality, end-to-end business
consulting and syndicated research services. Copyright © 2013 ACG Research. www.acgresearch.net.

Cisco Network Convergence System: Building the Foundation for the Internet of Everything

More Related Content

Similar to Cisco Network Convergence System: Building the Foundation for the Internet of Everything (20)

More from Cisco Service Provider (20)

Recently uploaded (20)

Cisco Network Convergence System: Building the Foundation for the Internet of Everything