Internet Protocol drives Universal Networking

Internet Protocol drives Universal Networking

Role of Internet Protocol in Networking

When digitalization of networks started taking place in the late eighties and  Integrated Services Digital Network (ISDN) was being developed for supporting data applications, the concept of OSI model came into being. OSI stands for ‘Open System Interconnection’ and the model classifies the entire end-to-end digital application as a set of seven distinct functions layered one on top of the other, each one entirely independent o f the others in terms of development and implementations. The model called for standardization of interfaces between the layers. The 7-layered OSI model is depicted in Figure 1.

Any application comprising of software is ordered in the form of packets of bit-wise information encapsulated in the presentation layer by appending a predefined leading header and possibly a trailing footer, that contain information and parameters pertaining to the presentation function. The presentation function in turn is encapsulated in the Sessions layer in similar manner, and so on until the process reaches the Physical layer. The physical layer defines the bits and bytes that get physically carried from origin to destination.

The Link Layer

The link layer is meant for transporting information between limited numbers of nodes with communication rules being defined at the link level. Here, a link is defined as a connection between a set of neighboring nodes. The communication rules are referred to as ‘protocols’. Initially, the link layer was conceived to serve a local area network within a building or a campus. Over a period of time, ‘Ethernet’ became a link layer protocol that gained tremendous popularity and virtually became a widely accepted standard. Employees in a company working in a common office could access their counterparts at speeds of 10Mbps or 100Mbps. The latter came to be known as the Fast Ethernet (FE). Later on, the concept of local area network was extended to the metro area within a city, where different commercial establishments such as banks or companies scattered at various geographic locations in the city had need for exclusive internal connectivity. The connectivity service provider soon had to contend with huge demand where he had to cater to providing metro area connections to thousands of establishments. There was therefore need to develop another link layer protocol that had to virtually carry the basic Ethernet between such customers. A Metro Ethernet Forum (MEF) was formed for the purpose of defining such services. To be able to deliver these services, the service provider needed to deploy serving nodes close to the customer premises. The function of these nodes would be to aggregate the traffic of different customers. There would then be a need to identify the user organizations. There would be need to control information queues to control traffic. All these set of new rules or protocols that enable these functions came to be known as the ‘Carrier Ethernet’. The new interfaces supported speeds of 1 Gigabit per second (Gbps) and then later on 10 Gbps, which came to be known as 1Gigbit Ethernet and 10Gigabit Ethernet respectively or simply as 1GE and 10GE.

The Network Layer

Consider a situation where one wants to step up one’s communications needs to the global level. In such a case, every node would need to have a global identity and if every such node needed to setup communication with any other node, a network level rule or protocol would be needed globally at the network level. The Internet protocol has turned out to be the most popular network layer protocol accepted by millions of Internet users. Like in the case of the Ethernet, here also, the service provider would need to setup aggregation nodes close to the customer locations that perform network functions such as routing the information packets to desired locations. The mechanism of maintaining route tables, updating them from time to time as per predefined rules and performing the routing functions are defined by standard protocols understood by every router. These nodes when interconnected together to provide any-node to any- node connectivity form a network of IP routers.

Initially, the Internet protocol called IPv4 or ‘IP version 4’ catered to only 2³² addresses which equal nearly 4 billion. Over the period of last two decades, it was realized by the Internet fraternity that this number was too small to cater to the needs of the growing number of Internet users. The Internet standards community called the IETF has subsequently upgraded the protocol to IPv6 or ‘IP version 6’ that caters to 2¹²⁸ addresses – a number which is sufficiently large to comprehensively cater to the needs of all the possible Internet users for a long time to come. The IPv6 protocol also includes a number of other new features that are likely to be needed in the future.

The Transport, Session & Presentation Layers

The Transport layer provides protocols that ensure reliable data transfer between users. The Sessions layer establishes, manages and terminates the connections between users, often referred to as Call control processes. The Presentation layer establishes context between Application Layer entities, in which the higher-layer entities may use different syntax and semantics if the presentation service provides a mapping between them. These layers are in most cases within the domain of the Telecom Service Provider.

The Application Layers

The Application layer interacts with software applications that implement a communicating component. Examples are Common Management Information protocol (CMIP), Hypertext Transfer protocol (HTTP), File Transfer protocol (FTP), Simple Mail Transfer protocol (SMTP) and Simple Network Management protocol (SNMP).

The Applications

On top of the Application layer, we have the specific Application which is visible to the user. The Application software works on the Computer operating system, and as such resides in the memory of the end terminal, say the PC, laptop or the mobile device. There are virtually thousands of applications that are used to drive the Input-Output devices connected to the computing system to deliver different type of services. The Input devices comprise of a variety of sensors such as optical scanners, pressure sensers, and so on, whereas the output devices range from printers, optical light emitting devices, sound producing devices, etc. The applications can be designed to interact with different devices spread anywhere in the network. Development of Applications therefore has become a multi-billion dollar industry where the developer only needs to know the computer programming language to develop the software without ever having to know the complexity of the intervening Telecom connectivity. Application development can therefore become a defining business for the IT professionals in India with enormous growth potential.

IP-MPLS Infrastructure

However, for being able to utilize these applications anywhere in the network, it is necessary to design and implement a network that will support the end user’s requirements in terms of the underlying layers that we described in the previous paragraphs. A Service Provider with an IP backbone may provide VPNs (Virtual Private Networks) which itself provides IP service to its customers.  MPLS (Multiprotocol Label Switching) is used for forwarding packets over the backbone. The BGP (Border Gateway Protocol) is used for distributing routes over the backbone.  The twin goals of this method are to support the outsourcing of IP backbone services for enterprise networks and for back-hauling mobile or broadband traffic. It does so in a manner, which is simple for the enterprise, while still scalable and flexible for the Service Provider, and while allowing the Service Provider to add value.

MPLS technology places labels on IP packets in a router. It categorizes or monitors the packets that traverse different routers in the network. MPLS is an overlay protocol MPLS is not designed to replace IP. Rather, it is designed as an overlay protocol that adds a set of rules to IP so that traffic can be classified, marked and policed.

MPLS-equipped networks use MPLS-aware devices known as label edge routers (LERs), positioned at the network’s edges. These devices are designed to inspect IP packets entering the network and add MPLS headers, as well as removing the headers from packets leaving the MPLS network. Inside the boundaries of the MPLS network, devices known as label switch routers (LSRs) look for an MPLS label on each packet that passes through them, looking up and following the instructions contained in those labels, routing them based on a list of instructions. Thus edge-to-edge Label Switched Paths (LSP) can be configured from one LER to another, through a series of LSRs, across the MPLS network. LSPs are pre-assigned and pre-engineered paths that packets with a certain label should follow without requiring the use of any dedicated lines such as the traditional SDH (Synchronous Digital Hierarchy) links. These are virtual circuits very similar to the circuit-switched paths in ATM or Frame Relay. One of the most obvious advantages of MPLS is that it provides network administrators with a number of tools for traffic engineering. An administrator, for example, can define a LSP that ensures VoIP traffic will be routed through the most reliable, highest performing sections of the network while less critical traffic, such as email, is sent across the slower sections.

Carrier Ethernet Aggregation Infrastructure

The IP Transport network is segregated into two parts –  (i) A Core Layer-3 IP-MPLS based network that covers major cities and hubs in the country connected through DWDM systems through the north side interfaces and (ii) An Aggregation Layer-2 Ethernet based network terminated at the IP-MPLS Edge node. An aggregation network as the name suggests, aggregates traffic over an area of say 100 km radius through a three tier architecture. The Hub of the aggregation network, called the Central Office Aggregation Unit (COAU) is collocated with the Next Generation Central Office (NG-CO) which houses the national level IP-MPLS Edge node.

The aggregation transport architecture in three tiers is indicated in the schematic in Figure 4.2. The COAU consists of the Tier-1 Carrier Ethernet (CE) switch which will handle traffic from STM-16 as well as the 10GE/100GE rings. The DWDM core is used to interconnect IP-MPLS core nodes, as well as directly interconnecting COAU Tier I switches / L2PE in case collocation is not possible. Tier-1 aggregation unit would generally utilize MPLS/RPR/100GE and would be planned for around 256 thousand users. The Tier-2 nodes are medium sized aggregation nodes called Next Generation Access Nodes (NG-AN) and would generally utilize MPLS/RPR/10GE with a plan for around 40 thousand users.

Tier- 3 nodes designated as Remote Access nodes are collocated with existing GSM towers and serve to provide backhaul for both mobile and fixed access traffic emanating within a radius of one to two kms. Enterprise traffic is also backhauled from Tier-3 nodes. The aggregation network could extend to neighbouring telecom centres through DWDM or other planned media. Tier-3 nodes would generally utilize MPLS/RPR/1GE with a plan for around 2 thousand users.

Synchronization standards for Ethernet called SyncE, defined in ITU-T G.8261, provides SDH-grade timing over lower cost 1GE and 10GE interfaces. This timing is important to ensure that time sensitive signals like voice and video do not suffer from slips beyond acceptable limits thereby preventing signal degradation. In WiMAX and LTE based mobile access that utilize OFDMA modulation techniques, phase synchronization becomes critical. In packet based networks, this can be achieved through providing an IEEE1588v2 clock.

The Next Generation Central Office at the Tier-1 location will house all the major functional systems as indicated in Figure-4.3.

Service deliverables

The integrated IP-MPLS and Carrier Ethernet Aggregation networks together would enable end-to-end MPLS services using FE, GE and 10GE interfaces from any node to any node in the country. Layer-2and Layer-3 VPN services would be available to enterprise customers. Backhaul services for mobile and broadband traffic as well as multicast services shall be available for the operator’s own use. The nodes would support classification of the traffic according to the port, VLAN, IEEE 802.1p bits or TOS/DSCP bits. Each classified Class of Service would be mapped to different EXP-bits in the MPLS header. In addition, there should be VLAN Tag support (IEEE 802.1Q) on the User Network Interface. These capabilities would allow the service provider to have reasonable degree of control over traffic engineering.

A schematic of the overall network comprising of a number of Layer-2 Carrier Ethernet-MPLS aggregation networks also referred to as ‘Converged Packet Aggregation Networks (CPAN) connected together through Layer-3 IP-MPLS network capable of delivering end-to-end MPLS services using FE, GE and 10GE interfaces is indicated in Figure 4.4.