Benefits, challenges and future trends of traffic management in different types of converged networks

Key Words: Network Traffic, Traffic Management, Converged Networks, quality of service, traffic control, policing, packet scheduling, priorities, queueing, buffers, packet discarding, congestion control, random early detection.

Introduction

General in computer networks are designed to handle a certain volume of traffic with an acceptable level of performance. Network performance deteriorate if the volume of network traffic exceeds the volume that can be handled by the network. There for long queues will be the order of the day

Traffic management is defined as a way of controlling how much traffic traverses each path in a network. This includes different mechanisms such as queueing, routing, restricting or rationing certain traffic on a network and giving priority to some types of traffic on certain network conditions

The goal of implementing all these mechanisms is to minimize the impact of congestion on the traffic’s quality of service. These can also be used to achieve certain performance goals and the proper implementation will ultimately improve the quality of an end-user experience technically or economically.

Several Traffic Management mechanisms are vital on a network that is traversing all sorts of Over the Top applications (Carela Español, 2014).

There are many Traffic Management mechanisms, such as

• Traffic Classification
• Traffic metering
• Traffic shaping
• Packet marking and or dropping
• Packet scheduling
• Admission control and resource reservation
• Routing decisions
• caching

Converged Networks is defined as an all in one where all types of network traffic are bundled together and transmitted over a single network. When data, voice, and video are transported across a network is identified by the network like any other application (Chandra and Kumar, 2015). The use of many communication modes on a single network has benefits of convenience and flexibility that are not possible with separate network infrastructures.

Converged networks contain a variety of traffic types that include voice and video, voice application, mission-critical, transactional, routing protocol and network management traffics. The requirement on the network differ depending on the mix of traffic types usually in terms of security and performance, for example, voice and video require constant bandwidth and low delay and jitter which is variation in delay (Wallace, 2011).

In a network, traffic that traverses through the network is heterogeneous and it flows from multiple applications and utilities. Most of these applications have a unique behavior and have their own requirements with respect to network parameters such as delays and jitter. These requirements must be met in order to avoid compromising the quality and usability of applications. Traffic management must exist in a network because it helps to prioritize different applications across the network with limited bandwidth and also ensuring that the requirements for each application have been met(Jose, 2008).

Network managers are expected to understand applications and protocols in network traffic because it is essential for the implementation of appropriate security policies. This enables various actions such as monitoring, discovery, and optimization to be performed on the identified traffic with the objective of improving network performance (Information, 2011).

Traffic Classification and Marking

Traffic classification and marking though are grouped together as one they are two distinct items within Quality of Service mechanisms to implement QoS. In general, their duty is to identify and split traffic into different classes and mark traffic according to the required behavior (Carroll, 2009).

Traffic Classification

Traffic classification is a tool that is used to sort packets into different traffic types so that different policies will be applied. This mechanism can be applied at every node in a network or it can be applied at the edge of the network were packets enter the network. Classification can happen without marking the packets. The mechanism inspects the packets to identify the type of traffic the packet is carrying. When the identification process has finished the traffic, types are handed to the treatment applications such marking, remarking and queuing (Chandra and Kumar, 2015).

There are a lot of traffic classification methods and approaches that used to group network traffic types, and these have gained relevance this decade. These classification ways are improving in accuracy and efficiency, but the continued proliferation of different application behavior in adding to disguise some applications to avoid filtering or blocking is among the reasons why traffic classification remains an open problem to research because new applications being developed using the same well-known ports to disguise their traffic and circumvent filtering or firewalls (Deb et al., 2013).

Despite the inaccuracy associated with the transport layer port with some application Traffic classification is still the fastest and simplest mechanism to continue monitoring traffic.

Traffic Marking

Traffic marking is also a traffic management mechanism that is used after the types of traffics have been identified. This is used in conjunction with traffic classification. It is defined as a process of modifying the attributes of traffic belonging to a specific category or class. Therefore it is the foundation for enabling many quality of service (QoS) features on the network. There are two methods of marking traffic which are

• using a set command
• and by creating a mapping table (Information, 2011).

Traffic marking mechanism has benefits to the network because it improves network performance by allowing to fine tune attributes of a network. This increase in granularity helps to single out traffic that requires special handling methods. Traffic marking also allows to segment a network into many priority levels (Chen, 2007).

This method is associated with many benefits that include allowing the administrator to determine how the network traffic is going to be treated based on how the attributes of network traffic are set. It allows the administrator to categorize network traffic based on many priority levels based on the attributes below(Information, 2011).

• Traffic is usually used to mark or set IP precedence or IP DSCP values for traffic entering into the network.
• It can be used to identify traffic for any class-based QoS feature.

Though traffic marking has benefits it also has challenges that affect the way it handles

Admission control is said to be the best and effective means of congestion control. The objective of this is to prevent congestion happening on a network. This is achieved by blocking new packet flows even before the flow begins. If a new flow traverses through the network, it means the network agrees explicitly or implicitly to provide quality of service through the duration of flow. Admission control methods are used to create new connections to decide whether the new connection will be accepted or rejected (Chen, 2007).

Admission control methods are based on the theory of probability and mathematical statistics and designed to keep the balance between the use of networks resources and the previously agreed connection parameters. These methods are the first protection against network redundancy.

Admission control methods have advantages that include allowing access to a maximum number of users and at the same time they are able to guarantee a good quality of service parameters. If we use AC methods to traffic control on many nodes in a network, it will increase the quality of service. Also using the same AC methods in all our nodes, it takes less time to make a decision as to which method to use in case each node makes the decision separately (Chen, 2007).

Although admission control is said to be the best in terms of traffic control. It comes with its limitations. Though this mechanism is said to be a natural approach, the complexity required for routers limits the scalability to large networks (Chen, 2007).

Traffic Policing and Shaping

Policing or shaping are mechanisms used to condition traffic before transmitting or when receiving traffic. These two mechanisms can work in tandem, but they are not mutually exclusive. Both mechanisms are used to measure the rate of different traffic classes of traffic against the traffic policy or a service layer agreement (SLA) set up by the administrator. This service layer agreement is usually set up between an enterprise and a service provider regarding bandwidth, traffic rate, reliability, and availability (Carroll, 2009).

Traffic Policing: Benefits

Bandwidth management through rate limiting.

This involves being allowed to control the maximum rate of traffic transmitted or received on an interface. It is usually configured on an interface at the edge of the network to limit inbound and outbound traffic to and from the network. In most traffic policing configurations traffic that falls within the parameters is transmitted but traffic outside the boundary limits will be dropped or transmitted with a different priority (Carroll, 2009).

Packet Marking

The benefits of traffic marking is that it allows a network administrator to partition there network into several priority levels or classes of service. The packets traversing through the network are marked so that the markings are used to identify and classify for downstream devices like ATM Cell Priority Marking or Frame Relay Discard Eligibility.

Packet marking is also used to set IP precidence or IP DSCp values for traffic getting into the network. The configure networking devices will then use the IP precedence marking values to decide on how traffic is supposed to be treated. This mechanism is also used to identify traffic for any type of class bsed QoS features. It can also be used to assign traffic to a QoS group in a device (Approach and Networking, n.d.).

Packet Prioritization for Frame Relay Frames is when the traffic policing feature enable users to mark Frame Relay DE bit of the Frame Relay Frame. The Frame Relay bit is only one, so it can be set to 0 or 1. In this scenario packets with a bit of 1 may be discarded first before the packets marked with a bit of 0.

Packet Prioritization for ATM Cells the traffic policy feature enables users to mark the ATM CLP bit in ATM Cells. The bit is used to prioritize packets in ATM networks.

Challenges

Traffic policing can be configured on an interface or sub interface.

Traffic policing is not supported on the EtherChannel interfaces.

Traffic Shaping

Traffic shaping is a mechanism that is used to manage the flow of data across the network and it helps to ensure network efficiency through maximizing bandwidth and congestion avoidance. Using this tool through configuring the system effectively will help to mange issues before an overload on the system. Traffic shaping regulates traffic by shaping it to a specified rate.

This tool uses a traffic descriptor that is indicated by classification of the packet to ensure that all the packet will adhere to the policies set to it and determine QoS to apply to every packet. Traffic shaping will help the administrator to control traffic leaving an interface and it matches its packet flow to the speed of a remote interface. Shaping a class of traffic to conform to downstream standards will help to eliminate bottlenecks in topologies with data rate mismatch (Mason et al., 2007).

There a lot of benefits that are gained by implementing traffic shaping mechanisms especially to the ISPs, classification gives the intangible but significant benefit of weeing what traffic passes through there network. In this scenario they can identify which subscribers are doing what on the network.

ISPs establish policies based on the IP or IP grouping of clients and ensure that users cannot defeat shaping by disguising protocols or encrypting their traffic. In addition to the above benefits intelligent shaping schemes also guarantees quality of service for a user at the same time allowing other traffic to use all remaining bandwidth (Deb et al., 2013).

Congestion Management

Congestion is defined as a situation where the rate of incoming traffic to an interface exceeds the rate of output from the interface. This happens usually when the network is connected to different Wide Area Networks, so traffic will be flowing in high volumes from a network interface to a low speed interface and the result of that is congestion. This is called speed mismatch problem. It also occurs if traffic from many interfaces aggregates into a single interface without enough capacity to handle the traffic, congestion is likely to happen, and this called the aggregation problem (Deb et al., 2013).

To solve the congestion problem several queuing methods were invented to perform congestion management. The introduction of these queuing methods was introduced as a benefit to solve congestion problems, but this is a temporary solution only.

There challenges associated with this queuing solution, because If the queue gets full, new arriving packets will be dropped and this is called tail drop. The queuing methods are based on a FIFO method where the packets that come first are delivered first and this pauses a lot of problem because many real time packets will be dropped.

Congestion Avoidance

This mechanism is used to avoid tail drop, that has several drawbacks. Conclusion. Congestion avoidance has two common techniques used on Cisco router interfaces. This is RED and its variations, namely WRED and CBWRED. This mechanism is one of QoS solutions (Akinade, 2015).

Packet Scheduling

Packet scheduling is a mechanism that address different needs of buffer packets scrambling for the same bandwidth link. It takes into account the requirements of each packet flow in quality of service. Packet flow requires a minimum input, maximum end to end delay or maximum packet delay jitter.

In packet scheduling the average delay time for all packets will not change under different schedules. This means giving priority to some packets at a different time will come at an expense of making packets is a network wait longer. The benefits that are gained for transmitting one packet flow disadvantages other packets (Zhang et al., 2014).

Conclusion

Traffic management is a very challenging problem as seen in the chapter that effort is being made in order to implement different mechanisms to control the flow of traffic in a network with a better quality of service. Many mechanisms exist but sometimes it is not enough to implement one or two or more at the same time at the same time not managing t solve the problems of congestion.

References

Akinade, A.O., 2015. Network Congestion Analysis and Control Management 2, 27–32.

Approach, A.E., Networking, C., n.d. Traffic management.

Carela Español, V., 2014. Network traffic classification : from theory to practice.

Carroll, B.J., 2009. Official Exam Certification Guide.

Chandra, P., Kumar, P., 2015. Internet Traffic Classification for Faster and Secured Network Service. Int. J. Comput. Appl. 131, 15–20. https://doi.org/10.5120/ijca2015907284

Chen, T.M., 2007. Network Traffic Management. Handb. Comput. Networks 1–40. https://doi.org/10.1002/9781118256107.ch22

Deb, R., Uddin, J., Islam, M., Islam, K.R., 2013. Traffic management Congestion Control Scheme in ATM Networks.

Information, F.F., 2011. Marking Network Traffic 1–26.

Jose, S., 2008. WAN and Application Optimization Solution Guide Cisco Validated Design.

Mason, L., Drwiega, T., Yan, J., Granelli, F., Kliazovich, D., Hui, J., Devetsikiotis, M., 2007. Managing Traffic Performance in Converged Networks 4516, 386-397–397. https://doi.org/10.1007/978-3-540-72990-7

Wallace, K., 2011. Implementing Cisco Unified Communications Voice over IP and QoS (CVOICE) Foundation Learing Guide, Pearson Education India.

Zhang, J., Chen, X., Member, S., Xiang, Y., Member, S., 2014. Robust Network Traf fi c Classi fi cation. Tatn 1–14.