Regardless of network’s age, the full integration of IPv6 services means that an inventory of the existing equipment must be performed to evaluate the cost of the project. This process would help plan the associated budget and project timescale. Although this step is mandatory to budget a deployment, it does not necessary help to evaluate the return on investment (RoI) associated with IPv6. In some cases, a given application or service clearly identifies a need for IPv6 and RoI becomes obvious. This leads to a rather targeted approach. When recognizing that the need for migration is inevitable, another way to integrate this new IP version is just to require products to be IPv6 capable for any new acquisition. In this way, the migration might occur over years as the networking environment gets upgraded to the newest generation of products and applications, but this planned approach will help to control the cost of the integration. The cost analysis includes the upgrade expenses for elements, such as hosts and network devices, but also labor for project planning and execution.

Host-Related Costs

An important requirement for enabling IPv6 on a host is the minimum release of its operating system that supports IPv6. This requirement extends to the capability for any application to run over an IPv6 network layer with full support from a given vendor. For example, Microsoft offered an IPv6 technology preview on Windows 2000, but it did not provide full support. Windows XP Service Pack 1 got a supported IPv6 stack but with a limited subset of supported applications such as Internet Explorer 6.0, Windows Media Player 9.0 and 10.0, and Conference XP 3.2, but no IPv6 support for popular applications such as Exchange, Outlook, or Microsoft Office. It is expected that the next-generation operating system, known as Microsoft Windows Vista, will deliver parity between IPv4 and IPv6 at the application level. Table 12-1 provides an overview of IPv6 stack availability on well-known operating systems at the time of this writing.

Table 12-1. IPv6 Support on Various Operating Systems

Vendor

Operating System

Reference

Apple

Mac OS X 10.2 and later

http://developer.apple.com/macosx/

BSD

FreeBSD 4.0 and later

OpenBSD 2.7 and later

NetBSD 1.5 and later

BSD/OS 4.2 and later

http://www.kame.net

HP

HP-UX 11i and later

Tru64 UNIX V5.1 and later

OpenVMS V5.1 and later

http://www.hp.com/products1/unix/operating/internet/ipv.html

http://h18000.www1.hp.com/ipv6/next_gen.html

IBM

AIX 4.3 and later

OS/390 V2R6 eNCS

z/OS Rel. 1.4 and later

http://www-306.ibm.com/software/os/zseries/ipv6/

Linux

RH 6.2 and later

Mandrake 8.0 and later

SuSE 7.1 and later

Debian 2.2 and later

http://www.bieringer.de/linux/IPv6/status/IPv6+Linux-status-distributions.html

Microsoft

Windows XP SP1 and later

Windows .NET Server 2003

Windows CE .NET (Pocket PC 4.1) and later

http://www.microsoft.com/ipv6

Novell

NetWare 6.1 and later

SUN

Solaris 8 and later

http://www.sun.com/software/solaris

Symbian

Symbian 7.0 and later

http://www.symbian.com

After the minimum releases for the targeted operating systems are identified, the cost of deployment will vary with respect to the hardware capability to support the respective software releases. This cost could go from zero in the case of hardware already running the appropriate version to significant amounts if the hardware must be replaced to run the latest operating system version. Table 12-2 provides a summary of the various possible scenarios.

Table 12-2. Estimate of Host Upgrade Costs

Hardware

Software

Minimal Operation

Cost

Full replacement

Full upgrade

Local intervention

Very high

Limited upgrade; for instance, memory size, disk, and so on

Full upgrade

Local intervention

High

No change

Full upgrade

Local intervention

Medium, depending on the licensing scheme

No change

Partial upgrade: for instance, service pack or application(s)

Remote intervention

High to minimal, depending on the licensing scheme

No change

No change, configuration only

Remote intervention

Low

The preceding process of evaluating the platform and operating systems should be followed by an inventory of applications. If the integration of IPv6 per application is considered to quickly take advantage of the protocol, it becomes simpler to classify the applications in categories that indicate the level of priority for a given corporate network (Table 12-3).

Table 12-3. Classification of Applications

Type of Applications

Priority

Comment

Application must be IPv6 capable

High

Minimum requirement to get IPv6 successfully running; for instance, DNS server and network management

Application should be IPv6 capable

High

Applications identified as business drivers to enable IPv6 in the enterprise

Application could be added if IPv6 capable

Medium

Applications identified as potential business drivers but not yet ready to run over an IPv6 network layer

Application might be IPv6 capable

Low

Applications that might migrate to IPv6 as the protocol becomes the incumbent

Costs associated with applications depend on their licensing scheme and the required upgrades. This can lead to significant costs when thousands of hosts running tens of applications are involved).

Network ElementsRelated Costs

It is likely that recently purchased routers and layer 3 switches are generally IPv6 capable. On the other hand, IPv6 support impacts all feature sets available on networking equipment, so a network manager needs to specifically list the services mandatory for his network to ensure their IPv6 equivalent is supported. In the case of Cisco IOS release trains, IPv6 features integrated over time are documented in Cisco IOS IPv6 Start Here manual (see http://www.cisco.com/ipv6).

Other networking devices such as layer 2 switches, wireless access points, firewalls, VPN concentrators, intrusion prevention system (IPS) might also require a certain level of IPv6 support to be compliant with the deployment guidelines. For this reason, the cost of upgrading nonrouter devices to support IPv6 should be included in the study, too.

Providing a generic network equipment industry status on IPv6 support might not be easy. You should instead poll your favorite vendors about the features of interest for your deployment. At the time of this writing, Table 12-4 offers an overview of IPv6 support in Cisco equipment.

Table 12-4. Cisco Equipment IPv6 Support

Type of Device

Status

Comment

Routers with IPv6 hardware forwarding capabilities

OK

Need to evaluate features such as filtering, multicast, and QoS to ensure they are supported.

Routers with IPv6 software forwarding capabilities

OK

Need to evaluate features such as filtering, multicast, and QoS to ensure they are supported.

L3 switches with IPv6 hardware forwarding capabilities

OK

Need to evaluate features such as filtering, multicast, and QoS to ensure they are supported.

L2 switches

OK

IPv6 traffic forwarding is transparent on layer 2 switch.

L2 switches

OK

Features such as MLD snooping impact the hardware.

Firewall

OK

Stateful and layer 7 filtering capabilities must be checked.

IPS

Under development

May require new hardware. Signature of attacks need to be computed for IPv6.

Encryption

OK

Generally require new hardware to support IPv6.

WiFi access point

OK

IPv6 traffic forwarding is generally transparent to WiFi AP.

WiFi access point

Under development

Management and security over an IPv6 network layer might not yet be available.

Storage

Under development

Might require hardware and software or just software upgrade.

Similarly to hosts, the cost of upgrading networking devices may vary from zero to significant amounts depending on the current level of readiness. Table 12-5 summarizes the expected cost of upgrading the network elements to support IPv6.

Table 12-5. Cost Estimate of Upgrading the Network Elements

Hardware

Software

Minimal Operation

Cost

Full replacement

Full upgrade

Local intervention

Very high

Limited upgrade; for instance, memory size, line card, supervisor engine

Full upgrade

Local intervention

High to Medium

No change

Full upgrade

Local or remote intervention

Medium to minimal, depending on the need to purchase an upgraded license

No change

No change, configuration only

Local or remote intervention

Minimal

Early planning for IPv6 deployment is critical to save some of these costs. Equipment-acquisition policies that mandate IPv6 support or a clear roadmap can lead to improved readiness with every scheduled network upgrade.

Operations-Related Costs

The integration of IPv6 in a production network represents a challenge for the operations team in charge of maintaining the integrity of IPv4 services. While working on the new protocol, they need to ensure seamless coexistence and similar service quality to the end users for both protocols. The project of integrating and running IPv6 has an evident impact on the budget of the department. Spending associated with training, design, product evaluation, service provider costs, deployment tasks such as product upgrade and configuration, and day-to-day operation, has to be planned and approved by management before the rollout can begin.

Configuration of hundreds or thousands of hosts and networking elements (independent of the need for hardware, operating system, or applications upgrade) is time-consuming and has a cost of labor that needs to be budgeted from day one. This must include the tasks associated with the update of the network management system and tasks related to the day-to-day support of the end users, the definition of updated operational processes, and their documentation.

Nevertheless, by planning for version-independent IP support through new acquisitions or development of applications, the cost of the upgrade could be integrated in the next rollout of equipment, which decreases the cost of deployment. Operating costs could also influence the selection of a given deployment strategy. Whether it mandates a deployment staging, tunnels before native or dual stack in the enterprise, or even leads to certain architecture choices (6PE versus native IPv6), operational cost could remain at its current level.