دانلود کتاب برش شبکه IPv6 – ارائه تجربه جدید برای صنایع

Reserving resources in networks to support or enable specific services is a well established practice dating back to circuit switch ing technologies that were in use even before the Internet was first put together. Connection oriented resource reservation became a critical component of transport networks built on the time division multiplexing network technologies SDH and SONET and was a fundamental element of ATM. These approaches, known as Traffic Engineering (TE), enabled operators to make service level guarantees and maximize the use of their network resources through the combination of policies applied to traffic entering the network, steering traffic onto specific paths within the net work, and management of the resources available to and consumed by the traffic as it transits the network. With the invention of Multiprotocol Label Switching (MPLS) at the end of the last century, it became possible to apply these traffic engineering techniques to packet based networks, and MPLS TE was born. An MPLS underlay pro vides tools for network operators to plan their packet networks, optimize the use of their resources, provide new services, and develop new revenue streams. Traffic engineering techniques enable network virtualization. A vir tual network provides the look and feel of a network built from physi cal resources and can be managed as a real network, but is constructed by dividing or partitioning the physical network resources to be shared by multiple virtual networks. Many network providers deliver virtual networks as services to their customers; Layer 2 and Layer 3 virtual pri vate networks (L2VPN and L3VPN) are the most popular examples. But, increasingly, more complex virtual networks are built and delivered for the customer to operate with their own resources. The Internet runs on IP, the Internet Protocol. IP is the glue that carries users’ data in packets across many interconnected networks based on a host of different technologies. For many years, everyone used IPv4, and everything was fine. However, it gradually became apparent that there would be more end systems attached to the Internet than there are globally unique addresses in the IPv4 address space, so the IETF worked to develop a new revision of IP, which was called IPv6. With 128 bit addresses, IPv6 had enough resources to address anything and everything. Although IPv6 had a slow start to its adoption, partly due to the ingenious invention of ways to extend the reach of the IPv4 address space, it is now very widely used, with around three quarters of a billion users connected to the Internet via IPv6 without even being aware of it. A nice feature of IPv6, beyond its large address space, is its ability to support mobile networking. The 3rd Generation Partnership Project (better known as the 3GPP), the primary standards organization for standardizing protocols for mobile communication, adopted IP as its data encapsulation and recognized transmission of mobile voice and data over the Internet as a fundamental part of the 3GPP architecture. Mobile communications have evolved from 3G and 4G to the latest 5G standards. Along with improvements in radio communications and the bandwidth connectivity that is supported, 5G has opened up aspirations for what the wireless user will be able to do with the service. Applications as diverse as holographic video conferencing and telesurgery, or self driving cars and enhanced online gaming, are considered to be on the horizon. These services place competing demands on the networks that support them. Some services may have a high priority (for example, public protection and disaster relief, or PPDR) with an absolute requirement that no data be lost. Other services may be highly sensitive to both data loss and variations in latency (such as remote surgery), while others may require the lowest possible latency (in cases varying from autonomous vehicles to automated trading to virtual or augmented reality gaming). Furthermore, some services will know a priori that they only need to exchange low levels of traffic, but others may require large, often very large, volumes. But the Internet was designed as a best effort delivery mechanism: every effort will be made to deliver data passed to the Internet, but it may be queued, re routed, buffered, and possibly dropped. This is not ideal for the delivery of these advanced 5G services, and additional mechanisms need to be developed to support them. A fundamental approach to meeting these requirements is known as Network Slicing. A network slice is a connectivity and resource commitment that models a VPN service with the assurance of meeting a set of specific network performance objectives. An IETF Network Slice is a network slice built using IETF technologies and expressed in terms of connectivity constructs and service objectives. Network slicing is a form of network virtualization and achieving network slicing with a traffic engineering technology such as MPLS is relatively straightforward. As an example, a set of TE tunnels provisioned with known properties and reserved resources could be considered a network slice. But building network slices on IPv6 is more complicated. It requires a combination of existing and new techniques in the data plane and the control/management plane. Fortunately, most of the building blocks already exist. We can call upon Differentiated Services (DiffServ), IPv6 extension headers, Segment Routing for IPv6 (SRv6), multi topology IGP, and control plane protocols like BGP LS and PCE, both of which I helped pioneer in the IETF. The few missing pieces are being actively worked on within the IETF. All that is needed is a cookbook to help the implementer know how to put all of the pieces together and build a functional network slicing system. This book is exactly that: an explanation of the protocols and encapsulations and an introduction to the network architectures and deployment modes that will enable IPv6 network slicing to support a wide range of applications and use cases. In many ways, network slicing is not a radical new idea. But its application to IPv6 networks is an innovation that proposes to offer significant opportunities for network operators to distinguish traffic flows and to deliver different treatments for traffic so as to guarantee the varied service levels that current and future use cases and applications will require. This will enable advances in user experience and the development of new technologies, which may open up new revenue streams for service providers.