IPv6 Trends in the Data Center

datacenterGuy3With IPv6 being mandated by the United States government, there has been a major push towards IPv6 recently in the US. Initially implemented by service providers, the mandate has forced anyone wanting to sell computing or networking hardware or software to the federal government to be able to support IPv6. Most data center operators realized that a failure to support IPv6 will eventually restrict access to connected resources and promote inefficient communications, not to mention limit potential customers that require IPv6. There are immediate benefits in transitioning to IPv6 in the form of improved network topologies and security.

Determining when and how to transition to IPv6 are common concerns. With the dwindling space availability of IPv4 addresses to support new devices, there is more pressure to move to IPv6, which has more benefits than IPv4. The upper layer application (except those with embedded Layer 3 addressing), typically has no visibility, which makes it difficult to detect whether it is running atop IPv4 or IPv6. Therefore, there tends to be limited or no changes to those applications to support IPv6. The applications that do have hooks into Layer 3 either have to be updated or replaced to support IPv6.

Most data centers today are building servers as virtual devices, which means that the number of physical devices is not representative of the number of servers that may be deployed within a given data center. The number of addressable devices may be much higher than what can be seen, leaving many data center operators to either be creative with their IPv4 addressing or deploy IPv6 sooner. Preexisting devices need to be transitioned to IPv6. The data center needs to be able to support clients from either stack, until there is a time that IPv6 is the prevalent Layer 3 protocol.

IPv6 Transition Strategy
On July 14, 2013, the Internet Engineering Task Force (IETF)’s RFC draft on IPv6 Operation Guidelines for Datacenters stated:

“Three main transition stages can be considered when analyzing IPv6 deployment in the DC infrastructure, all compatible with the availability of services running in the DC through IPv6:

  1. Experimental. The DC keeps a native IPv4 infrastructure, with gateway routers (or even application gateways when required by services) performing the adaptation to requests arriving from the IPv6 Internet.
  2. Dual stack. Native IPv6 and IPv4 are present in the infrastructure, up to whatever the layer in the interconnection scheme where L3 is applied to packet forwarding.
  3. IPv6-Only. The DC has a fully pervasive IPv6 infrastructure, including full IPv6 hypervisors, which perform the appropriate tunneling or NAT if required by internal applications running IPv4.”

There is no clear date for when IPv6 has to take over IPv4. The mandate identifies a date for native support of IPv6 but not for disabling IPv4. As an interim solution, running dual stack on devices seems like a reasonable solution while maintaining backward compatibility with legacy IPv4 systems. All devices along the communication path, including endpoint devices, data center routers and switches, ISPs, and backbone providers must simultaneously support IPv4 and IPv6.

Although the dual-stack model may appear to add complexity and the need for additional bandwidth, it has the benefit of ensuring that no matter which protocol a device’s application uses, there will be support for it in the underlying network topology. Nearly all current operating systems support both IPv4 and IPv6, and so frequently, enabling IPv6 is simply a matter of turning it on (or just making sure that no one has turned it off).

Once dual-stack operating systems are deployed and enabled throughout a data center, there are a few tunneling technologies (6in4, 6to4, ISATAP, and Teredo) that ensure that IPv6 traffic can traverse through a portion of the network where only IPv4 is supported.

The transition to IPv6 takes time, effort, and careful planning. The allocation of addresses needs to be carefully considered in order to make IPv6 manageable and scalable.

The final stage of transitioning to IPv6 will be native support of IPv6 on all devices and disabling IPv4 support (or supporting IPv4 though NAT64). There are several motivations to transition to a native IPv6-only data center. One compelling motivation is cost and supportability. With a single protocol, we can reduce the cost of support and possibly reduce the number of components needed within the data center.

Modern data center networks need to keep pace with advancing technology in order to deal with the demands of their clients and traffic. With the transition to IPv6 within enterprises, governments, and service providers, the data center networks have to keep up with growth (or in some cases, lead the way). For most applications, there are no changes. However, the infrastructure must be in place in order to support the needs of those applications, which is not a hypothetical or future-oriented scenario—it is happening now and we need to be involved.

Related Courses
IPv6FD – IPv6 Fundamentals, Design, and Deployment v3.0
DCUFI – Implementing Cisco Data Center Unified Fabric v5.0
DCICN – Introducing Cisco Data Center Networking v1.0

In this article

Join the Conversation