SRv6 is coming

Segment Routing MPLS customer adoption is truly outstanding. As of today, we count more than 30 live deployments and about 80 deployments are planned.

The pace of adoption for SRv6 will be even faster. Let me expand on why I’m making this bet.

IPv6 connectivity is getting more pervasive

Back in 2012, major Internet service providers, home networking equipment manufacturers, and web companies around the world made the common decision to enable IPv6 for their products and services. Since then, IPv6 adoption has been steadily increasing, but at a much lower growth rate than expected. The upcoming fourth industrial revolution fueled by 5G connectivity is set to change the game and accelerate IPv6 deployments.

As Asia, Europe, North America, and Latin America have already exhausted their IPv4 allotments, and Africa is expected to exhaust its allotment by 2019, the transition from an IPv4 environment to an IPv6 environment is becoming of utmost importance.

IPv6 is the new normal and many service providers are actively engaged in upgrading their network infrastructures.

SRv6 inherits all of SR MPLS capabilities

Simplification

SRv6 further simplifies the network by eliminating MPLS altogether and by relying on the native IPv6 header and header extension to provide the same services and flexibility as SR-MPLS, directly over the IPv6 data plane.

Resiliency

Resiliency plays a pivotal role in ensuring the network stays up always so that customers can access their services from anywhere at any time.

Current routing protocols – IS-IS, OSPF – provide a first level of resiliency by rerouting traffic around failures in the network. But it’s not enough, more and more applications need the network to guarantee under 50ms protection against any kind of network failures. This is exactly what SRv6 TI-LFA (Topology Independent Loop Free Alternate) brings with 100% topology coverage, simplicity and path optimality.

Traffic Engineering

Leveraging the most advanced SRv6 traffic engineering capabilities, the network can be turned into a multi-service infrastructure. New Flexible Algorithm capabilities make multiple optimizations of the same physical network infrastructure along various dimensions possible (e.g., one can be optimized for low-latency vs. another one for bandwidth, or one can offer disjoint paths via two distinct planes.)

Network slicing will play a major role as service providers and enterprises get ready to offer a wide range of 5G services, that have specific and differentiated needs, over a converged infrastructure. As a result, service providers are implementing top-notch traffic engineering solutions across their network, directly from the cell site and up to the core and data centers, to ensure each service gets its own dedicated networking slice with its own set of SLAs.

SRv6 adds network programming capabilities

SRv6 takes advantage of IPv6 Extension Headers by inserting Segment Routing headers into IPv6 packets. Any IPv6 packet can now contain a list of Segment Identifiers (Segment IDs), that are nothing else than 128-bits IPv6 addresses. Thanks to the increase in Segment ID size, it is now possible to pack more than mere IP addresses into a Segment ID and hence go beyond routing purposes.

This opens the door to infrastructure programming. The first 64 bits can be used to direct traffic to a specific node in the network – the “main body” of the program – the next 32 bits can be used to enforce some actions on the traffic – the “function” part – and the remaining 32 bits can be used to pass some additional information – the “argument” part.

Supercharging the network with these programming capabilities is a game-changer in the way the network treats applications. Your network is no longer merely routing traffic from point A to point B according to some specific constraints expressed by applications (e.g., SR traffic engineering). The network can now take actions on the applications along the same path applications are transported over. It’s about making your applications and your network interact in a completely different, new way.

SRv6 is enjoying strong adoption

SRv6 early adoption has been remarkable with some noteworthy live deployments supporting significant customer traffic:

• SoftBank Corporation aims to further enhance the efficiency and functionality of its network by introducing the latest technologies such as SRv6 with the intent to roll out a highly reliable mobile network that can cope with the future traffic for the age of 5G and IoT.
• Iliad: The SRv6 deployment enables Iliad to build a network that is extremely scalable with improved reliability, flexibility, and agility, all while helping to reduce CapEx and OpEx. To further expand the SRv6 benefits across the entire network, Iliad has developed its own SRv6 software stack that will equip its homegrown “NodeBox” designed to aggregate the traffic from mobile base stations.
• LINE Corporation uses an SRv6 overlay to provide per-service policy on a shared underlay network in the data center. The hypervisor takes care of encap/decap, and OpenStack is used as the SRv6 control plane with in-house developed extensions.

SRv6 ecosystem is growing

Any new technology can only make inroads into service providers’ networks when it is supported by a rich ecosystem; otherwise, the risks associated with rolling out a new technology are simply too high.

SRv6 is benefiting from a rich ecosystem spanning across:

• Networking vendors – Nokia, Huawei
• Silicon chipset manufacturers – Barefoot, Broadcom, Marvell
• SmartNic manufacturers – Intel
• Open source – Fd.io, Snort…

It does not stop there. In the next coming months, you will hear about more companies taking advantage of these opportunities.

SRv6 standardization is well on its way

A key milestone was achieved in October 2019 with the Segment Routing Extension Header (SRH) draft proposal that became a proposed standard.

Another key proposal is SRv6 Network Programming. While it is still in draft, it is in its advanced stages and could become a proposed standard as early as March 2020.

Finally, new SRv6 network programming capabilities, such as the SRv6 uSID instruction, have been posted.  They provide for ultra large-scale deployments (multi-domain networks with 100k routing nodes or datacenters with billions of servers) and leverage legacy equipment.

Over the next coming weeks, you will hear from some industry leaders who will share their thoughts on SRv6 – why they believe it is a game changer, and what their plans are.

In the meantime, if you want to learn more about SRv6, visit us at segment-routing.net.

Read a comprehensive update on the
current state of SRv6 standardization.