Skip to main content

What is QoS?

Quality of service (QoS) is the term used to describe or quantify the overall performance of a service, such as a telephone network, or a cloud computing service, and more specifically the performance perceived by users. In the field of computer networks, Quality of Service (QoS) is a collection of technologies that operate together on a network to ensure that it can reliably execute high-priority applications and traffic even when network capacity is restricted. This is accomplished via the use of QoS technologies, which provide differentiated handling and capacity allocation for various flows of network traffic. This allows the network administrator to control the sequence in which packets are processed and the amount of bandwidth allocated to a particular application or traffic flow.

Latency (delay), jitter (variance in latency), bandwidth (throughput), and error rate are all metrics that affect QoS. This emphasizes the necessity of QoS for high-bandwidth, real-time traffic such as video conferencing, voice over IP (VoIP), and video-on-demand, which are particularly sensitive to latency and jitter.

Nowadays, businesses are expected to deliver trustworthy services with little disturbance to the end-user. In recent years, as applications such as audio, video, file sharing, and streaming data have become more integrated into our daily lives, QoS has become more critical.

Due to the rising number of connected devices, the amount of application usage, and the huge increase in social media traffic, a network may easily get saturated. This network saturation might result in performance disparities. As a consequence, IT teams get a flood of complaints of interrupted video meetings, poor audio quality, delays, and even missed phone conversations, all of which may dramatically impair day-to-day workplace efficiency.

Quality of service is vital for any business that wishes to ensure the optimal functioning of its mission-critical apps and services. It is critical to avoid latency or lag in high-bandwidth solutions such as videoconferencing, VoIP, and, increasingly, streaming services.

QoS allows an organization to prioritize traffic and resources to ensure that a certain application or service performs as promised. Additionally, it allows organizations to prioritize various applications, data flows, and users to ensure their networks operate at peak performance.

We will discuss in detail what QoS is in computer networks, how it works, the benefits of QoS, QoS applications and service models, and the history of QoS in this article. Additionally, we will explore the top Quality of Service software available on the market.

How Does QoS Work?

QoS networking technology works by identifying service types in packets and then configuring routers to establish different virtual queues for each application depending on their priority. As a consequence, bandwidth is allocated for mission-critical applications or websites for whom priority access has been granted.

Quality of service (QoS) systems allocate capacity and handling to specified flows of network traffic. This allows the network administrator to prioritize packet processing and provide the proper amount of bandwidth to each application or traffic flow.

How QoS network software operates is contingent upon specifying the different sorts of traffic that it monitors. These include the following:

  • Bandwidth: The rate at which a connection transmits data. QoS may instruct a router on how to allocate bandwidth. For instance, allocating a certain amount of bandwidth to distinct queues for distinct sorts of traffic.
  • Jitter: The erratic pace at which packets travel across a network as a consequence of congestion, which may result in packets arriving late and out of order. This might result in distortion or gaps in the supplied audio and video.
  • Loss: The quantity of data lost due to packet loss, which is most often caused by network congestion. Organizations may use QoS to choose which packets to discard in this circumstance.
  • Delay: The amount of time required for a packet to get from its origin to its destination. This is often impacted by queuing delay, which happens when a packet is held in a line before being broadcast during periods of congestion. By establishing a priority queue for particular kinds of traffic, QoS helps businesses to prevent this.

To implement QoS, a company must first identify the kinds of traffic that are critical to them, use a large amount of bandwidth, and/or are sensitive to delay or packet loss. This enables the firm to comprehend the requirements and significance of each sort of traffic on its network and to develop a comprehensive strategy. For example, some businesses may simply need configuring bandwidth restrictions for certain services, whilst others may require configuring interface and security policy bandwidth limits for all their services, as well as prioritizing key services based on traffic rate.

After that, the firm may implement rules that categorize traffic and maintain the availability and consistency of its critical applications. Classification of traffic may be done by port or internet protocol (IP), or by a more advanced method such as application or user.

After that, bandwidth management and queuing tools are given duties to regulate traffic flow according to the categorization they acquired upon network entry. This enables the storage of packets inside traffic flows until the network is prepared to handle them. Priority queuing may also be used to guarantee the network's performance is available and has a low latency for critical applications and traffic. This is to ensure that the network's most critical operations are not deprived of bandwidth by lower-priority activities.

Additionally, bandwidth management monitors and manages network traffic flow to ensure that it does not exceed capacity or cause congestion. This comprises traffic shaping, a rate-limiting approach that improves or ensures performance while increasing available bandwidth, and scheduling algorithms, which provide many strategies for allocating bandwidth to particular traffic flows.

Why is QoS Important?

Historically, corporate networks functioned as distinct entities. One network handled phone calls and teleconferences, while another handled laptops, workstations, servers, and other devices. They seldom crossed paths, unless a computer was connected to the internet through a telephone connection. When networks transported merely data, speed was not a priority. However, interactive apps with audio and video material must now be sent at a high rate, without packet loss or delivery speed changes.

QoS is vital for ensuring the excellent performance of mission-critical applications that demand a large amount of bandwidth for real-time traffic. For example, it enables enterprises to prioritize the performance of "inelastic" services like VoIP and videoconferencing that often have low bandwidth needs, strict latency constraints, and high sensitivity to jitter and delay.

QoS enables organizations to avoid delays in these critical applications, ensuring they function at the level required by users. For example, missing packets might create a delay in the stream, resulting in choppy and indecipherable voice and visual quality during a videoconference session.

Quality of service (QoS) is becoming more crucial as network performance needs to adjust to the rising number of users. Modern apps and services need enormous amounts of bandwidth and network performance, and consumers want them to operate at peak speed at all times. As a result, organizations must use procedures and technology that provide the greatest possible service.

Quality of service (QoS) is also becoming more crucial as the Internet of Things (IoT) matures. For instance, in the industry, machines now communicate with one another over networks to deliver real-time status reports on any possible problems. As a result, any delay in reporting might result in very expensive errors in IoT networking. QoS allows the data stream to get network priority and guarantees that information flows as rapidly as feasible.

Cities are increasingly densely packed with smart sensors, which are critical for the successful operation of large-scale IoT initiatives such as smart buildings. Data gathered and evaluated, such as humidity and temperature readings, is often time-sensitive and must be carefully recognized, annotated, and queued.

What are the Advantages of QoS?

QoS is vital for firms seeking to assure the availability of mission-critical applications. It is critical for providing differentiated bandwidth and ensuring that data is sent without interfering with traffic flow or resulting in packet loss. Several significant benefits of implementing QoS include the following:

  • Reduced Latency: Latency is the time required for a network request to travel from the sender to the receiver and be processed by the receiver. This is often impacted by the greater time required for routers to examine data and storage delays introduced by intermediary switches and bridges. In an ideal world, the delay of these packets should be zero. Even yet, if a delay occurs, it might result in overlapping audio in IP audio and speech packets or an echo effect for the receiver. If real-time transport protocol (RTP) packets are left unclassified, network delay may be a prevalent and frustrating issue for IT organizations. Classification and prioritization are critical in these situations to minimize concerns with delays in video and audio IP transfers. By prioritizing their key applications, QoS helps enterprises to minimize latency and accelerate the processing of network requests.
  • Reduced Packet Loss Rate: Packet loss occurs when data packets are lost in transit between networks. This is often caused by network congestion, a broken router, a loose connection, or a bad signal. By prioritizing bandwidth for high-performance applications, QoS eliminates the possibility of packet loss.
  • Prioritizing the Applications: The primary advantage of Quality of Service is that it may help an organization's IT infrastructure enhance the performance of important applications. By identifying and prioritizing critical applications based on their network traffic, QoS may assist guarantee that critical apps meet their packet loss, delay, and latency requirements. QoS ensures that enterprises' most mission-critical applications are always given precedence and the resources required to function well.
  • Better User Experience: The ultimate purpose of QoS is to ensure the high performance of mission-critical applications, which equates to an ideal user experience. Employees benefit from great performance on their high-bandwidth programs, which helps them to be more productive and complete tasks faster.
  • Improved Resource Management: The majority of firms have invested significantly in their network infrastructure and maybe employ more costly transport media. For instance, a company may have MPLS lines established or may rely on mobile networks for redundancy and resilience. Since QoS network traffic identification may prioritize traffic depending on the associated application, application-specific rules can be developed to guarantee that costly and high-performance network bandwidth is utilized mostly or exclusively by the applications that demand it. QoS helps managers to better control the internet resources available to their firms. Additionally, this decreases expenditures and the requirement for connection expansion investments.
  • Management of Point-to-Point Traffic: Managing a network is critical regardless of how traffic is provided, whether it is end-to-end, node-to-node, or point-to-point. The latter allows enterprises to transport client packets in order over the internet without experiencing packet loss.
  • Enhanced Security: QoS has the power to filter out undesirable or suspect data traffic that crosses its route, thereby functioning as a firewall and forming an integral part of more secure network architecture. Additionally, security regulations mandate that encrypted packets be prioritized, ensuring that only secure data packets are sent.
  • Optimized Traffic Routing: Because traffic from various applications has distinct destinations, a one-size-fits-all approach to traffic routing might result in inefficiencies and delays. An organization may more effectively route network traffic to its destination by identifying the application associated with a given network connection and applying application-specific rules to it.

What is QoS in Networking?

Quality of service (QoS) is a term that refers to the use of procedures or technologies on a network to manage traffic and assure the functioning of mission-critical applications when network capacity is constrained. It lets enterprises manage their network traffic more efficiently by prioritizing high-performance apps.

Typically, QoS is used in networks that transport data for resource-intensive systems. It is often needed for services such as internet protocol television (IPTV), video on demand (VOD), voice over IP (VoIP), online gaming, and videoconferencing.

By implementing QoS in networking, enterprises may improve the performance of different applications and obtain insight into their network's bit rate, latency, jitter, and packet rate. This enables them to engineer network traffic and alter the way packets are routed to the internet or other networks to avoid transmission delays. Additionally, this guarantees that the company meets the anticipated service quality for apps and provides the anticipated user experience.

According to the QoS definition, the primary objective is to allow networks and organizations to prioritize traffic via the provision of dedicated bandwidth, controlled jitter, and low latency. This is critical for optimizing the performance of corporate applications, wide-area networks (WANs), and service provider networks.

How to Identify QoS in the Network?

Businesses may use a variety of approaches to ensure the excellent performance of their mission-critical apps. These include the following:

  • Queuing: Queuing is the process of establishing rules that give some data streams priority treatment over others. In routers and switches, queues are high-performance memory buffers in which packets traveling through are stored in designated memory sections. When a packet is given a higher priority, it is transferred to a specialized queue that pushes data more quickly, reducing the likelihood of it being lost. For instance, corporations may implement a strategy that prioritizes voice traffic over the bulk of network capacity. The routing or switching device will then transfer the packets and frames associated with this traffic to the front of the queue and transmit them instantly.
  • Resource Reservation: The Resource Reservation Protocol (RSVP) is a transport layer protocol that reserves resources across a network and may be used to provide application data streams with a defined degree of QoS. Businesses may use resource reservation to segment network resources according to the types and origins of traffic, impose limitations, and guarantee bandwidth.
  • Prioritization of Time-Sensitive Traffic: Numerous corporate networks may become overcrowded, causing routers and switches to discard packets as they arrive and depart quicker than they can be handled. As a consequence, streaming services are adversely affected. Prioritization permits traffic to be categorized and assigned a different priority based on the type and destination of the traffic. This is especially advantageous in situations of heavy congestion, since higher-priority packets may be transmitted ahead of other traffic.
  • Traffic Marking: Once it is determined which applications demand precedence over other capacities on a network, the traffic must be tagged. This is accomplished using mechanisms such as Class of Service (CoS), which indicates the presence of a data stream in the Layer 2 frame header, and Differentiated Services Code Point (DSCP), which indicates the presence of a data stream in the Layer 3 packet header.

Along with these strategies, companies should bear numerous best practices in mind while defining their QoS needs.

  • Consider the distribution of packets between available queues and the queues that are utilized by which services. This may affect the latency of the network, the distribution of queues, and the assignment of packets.
  • Prioritize all traffic using either service-based prioritization or security policy prioritization, but not both. This simplifies the analysis and troubleshooting process.
  • Only particular services should be guaranteed bandwidth. This prevents all traffic from sharing the same queue in high-volume scenarios.
  • Ascertain that the source interface's maximum bandwidth limitations and security policy are not set too low to avoid excessive packet rejection.
  • Utilize the User Datagram Protocol (UDP) for reliable testing results and avoid oversubscribing bandwidth capacity. To guarantee good performance, attempt to keep the complexity of the QoS setting to a minimum.

What are QoS Applications?

Today's interactive applications that transport audio and video must be provided over networks at fast rates and without packet loss or delivery speed changes. Individuals now conduct business conversations using video-conferencing software such as GoToMeeting, Zoom, and, Skype which makes use of the IP transport protocol to transmit and receive video and audio data.

Certain forms of network traffic that may need or demand a particular level of service are as follows:

  • Streaming media especially; IPTV, Ethernet-based audio, IP-based audio

  • Voice over Internet Protocol (VoIP): VoIP requires a minimal level of packet loss, delay, and jitter. If these standards are not satisfied, both callers and receivers will experience poor call quality. To remedy this issue, adjust priority mapping such that voice packets take precedence over video packets and traffic policing so that voice packets get the maximum available bandwidth. This guarantees that voice packets are prioritized during network congestion.

  • Telepresence

  • Videotelephony: Video conferences need a large amount of bandwidth, a short latency, and little jitter. Configure traffic policing to provide sufficient bandwidth for video packets and priority mapping to boost the priority of video packets to fulfill the needs of such services.

  • Industrial control systems protocols such as EtherNet/IP are used to control machines in real-time.

  • Online games with the potential for real-time latency

  • Storage technologies such as iSCSI and Fibre Channel over Ethernet

  • Safety-critical applications, such as remote surgery, might be dangerous due to availability difficulties.

  • Support systems for network operations, either for the network itself or for clients' mission-critical requirements

  • Networking and administration Protocols, such as OSPF and Telnet: These services demand minimal latency and packet loss, but do not require a large amount of bandwidth. To address the needs of such services, enable priority mapping to map the service packets' priority to a higher CoS value, allowing the network device to transmit the packets preferentially.

These applications, which need a minimum amount of bandwidth and have a maximum latency limit, are referred to as "inelastic."

Heavy-traffic services that include the transmission of a huge volume of data over an extended period, such as FTP, database backup, and file dump, need a low packet loss rate. You should configure traffic shaping to cache service packets transmitted from an interface in the data buffer to fulfill the needs of such services. This minimizes packet loss under burst traffic congestion.

Some common services, such as HTML web page surfing and email do not have any network-specific needs. You are not required to implement QoS for them.

What are QoS Service Models?

To implement QoS, three approaches are available: Best Effort, Integrated Services, and Differentiated Services.

  1. Best Effort: A QoS approach in which all packets are treated equally and there is no assurance of packet delivery. When networks do not have QoS rules specified or when the infrastructure does not support QoS, Best Effort is used. It is a less often employed model, although it is by far the most simple. Best Effort is often used as a default model when networks do not yet have any QoS regulations specified.
  1. Integrated Services (IntServ): A quality of service approach that reserves bandwidth along a specified network route. Applications request resource reservations from the network, and network devices monitor the flow of packets to ensure that network resources are available to receive the packets. IntServ implementation needs IntServ-capable routers and network resource reservation through the Resource Reservation Protocol (RSVP). Network devices may analyze the flow of data packets and guarantee that adequate space is available to receive the required packets. IntServ's scalability is restricted, and it consumes a large number of network resources.
  1. Differentiated Services (DiffServ): DiffServ, perhaps the most widely used QoS model, is a concept of quality of service in which network devices such as routers and switches are set to support several types of traffic with varying priority. Network traffic must be classified according to a company's setup. Administrators provide a DSCP (Differentiated Services Code Point) value between 0 and 63 to each kind of traffic to prioritize it and organize it according to traffic classifications (TCs). These values may be specified in the appropriate headers. Data with a high DSCP value will be prioritized and will reach their intended destination without delay or disturbance. DiffServ fully utilizes the flexibility and extensibility of IP networks by converting information included in packets into per-hop behaviors (PHBs), significantly decreasing signaling operations. The DiffServ approach incorporates the following Quality of Service (QoS) mechanisms:
  • Classification and marking: Classification and marking of traffic are necessary conditions for offering differentiated services. Traffic classification categorizes packets and may be accomplished via the use of traffic classifiers programmed using the Modular QoS Command-Line Interface (MQC). Traffic marking gives packets distinct priorities and may be achieved through priority mapping and re-marking. Packets include a variety of various sorts of precedence fields, which vary according to the network type. In a VLAN network, for example, packets carry the 802.1p field; in an MPLS network, they carry the EXP field; and in an IP network, they carry the DSCP field.
  • Traffic policing, traffic shaping, and rate-limiting at the interface: Traffic policing and traffic shaping are used to keep the traffic flow within a certain bandwidth. When the traffic rate exceeds the limit, traffic policing reduces excessive traffic, while traffic shaping buffers excess traffic. On an interface, traffic policing and traffic shaping may be used to create interface-based rate limitations.
  • Management of congestion and congestion avoidance: Congestion management queues packets and sets the forwarding order using a particular scheduling algorithm when the network is congested. Congestion avoidance monitors network resource utilization and removes packets when congestion develops to prevent network overload.

What are the Best QoS Tools?

To maximize the advantages of QoS models and procedures for prioritizing network traffic, system administrators must constantly monitor their network's health. Network performance tools are critical for administrators wishing to more effectively monitor and prioritize network traffic inside their organization. Network performance analyzers give IT teams a full picture of their network traffic in real-time, enabling them to immediately identify problems and take proactive actions to avoid recurrences. Generally, QoS tools fall into the following categories:

  • Queueing: Reserves bandwidth for the storage of packets in a buffer for later processing.
  • Policing: Enforces a certain bandwidth allocation and limits or removes packets that violate the regulation. This is a method of avoiding congestion.
  • Shaping: Similar to policing, it queues excess traffic in a buffer rather than dropping it entirely. This, along with queueing, is a method of managing congestion.
  • Classification: Identifies and tags traffic to ensure that it may be identified and prioritized by other network devices.
  • Random early discard with a weighted distribution (WRED): Reduces the priority of low-priority data flows to protect high-priority data from the adverse consequences of network congestion.
  • Compression and Fragmentation: Reduces the bandwidth available on a network to eliminate latency and jitter.

QoS tools may accomplish all of these tasks or a subset of them. Numerous Quality of Service tools is controlled and execute these activities automatically, allowing administrators to modify settings and rules as necessary. Several common quality of service monitoring tools include the following:

  • ntopng
  • ManageEngine NetFlow Analyzer
  • Solarwinds NetFlow Traffic Analyzer

1. ntopng

ntopng is the successor to ntop, a network traffic probe that observes network activity. ntopng is built on libpcap/PF RING and was designed to be portable across all Unix platforms, macOS, and Windows. It is a simple, encrypted online user interface for exploring real-time and historical traffic data.

Ntopng is an excellent choice for small businesses seeking a straightforward, open-source Quality of Service application. While its simple dashboard is enticing, it also enables managers to monitor traffic by IP address, port, and across the network, providing extensive statistics and simplifying network planning.

Principal Characteristics of ntopng are as follows:

  • Display network traffic and active hosts in real-time
  • Classify network traffic based on a variety of characteristics, including IP address, port, Layer-7 (L7) application protocols, throughput, and Autonomous Systems (ASs)
  • Top talkers (senders/receivers), top application servers, and top L7 application protocols
  • Analyze IP traffic and classify it by source/destination
  • Report Usage of IP protocols classified by protocol type
  • Utilize nDPI, ntop Deep Packet Inspection (DPI) technology to discover Layer-7 application protocols ( YouTube, Facebook, BitTorrent, and others).
  • Produce long-term reports on a variety of network parameters, such as throughput and L7 application protocols.
  • Monitor and report real throughput, network and application latencies, Round Trip Time (RTT), TCP statistics (retransmissions, out-of-order packets, and packet loss), as well as bytes and packets transferred.
  • Persistently save traffic information on disk to facilitate future investigations and post-mortem studies.
  • Support for exporting monitored data to MySQL, ClickHouse, and ElasticSearch
  • Interactive historical study of monitored data exported to ClickHouse
  • Support for SNMP v1/v2c/v3 and continuous monitoring of SNMP devices
  • Management of identities, including the association of VPN users with traffic
  • Geographically find and overlay hosts on a map
  • Statistical analysis of HTML5/AJAX network traffic
  • IPv4 and IPv6 are fully supported.
  • Support for Layer-2 in its entirety (including ARP statistics)
  • REST API to facilitate integrations with third-party applications
  • Detunnelling of GTP/GRE
  • Analyses of behavioral traffic, such as lateral movements and identification of periodic traffic

ntopng Dashboard

Figure 1. ntopng Dashboard

2. ManageEngine NetFlow Analyzer

ManageEngine NetFlow Analyzer is a comprehensive traffic analysis tool that makes use of flow technologies to give real-time insight into network bandwidth performance. This comprehensive program is used for network forensics, application monitoring, capacity planning, and bandwidth capacity analysis. It has the following features:

  • Supports a variety of common flow technologies and devices from well-known suppliers.
  • Provides real-time insight into your network's traffic and bandwidth.
  • Assists in configuring real-time threshold-based alerts to shorten reaction times and facilitate troubleshooting.
  • Monitors network traffic patterns in a proactive manner in order to identify traffic spikes and abnormalities.
  • Internal and external security risks such as DDoS/flash crowd attacks, probes/scans, and unusual flows are detected.
  • Allows you to forecast, plan, and optimize bandwidth utilization in order to guarantee that business-critical applications get priority.
  • Contributes to the analysis of bandwidth utilization patterns in order to pinpoint the source of network traffic problems.

ManageEngine NetFlow Analyzer Dashboard

Figure 2. ManageEngine NetFlow Analyzer Dashboard

3. SolarWinds NetFlow Traffic Analyzer

Network administrators will be able to resolve issues using the Netflow Traffic Analyzer. It helps them with providing QoS implementation and optimization, monitoring bandwidth utilization in order to determine which programs and devices are hogging network resources.

This fashionable Quality of Service application provides CBQoS 'class-based quality of service' monitoring, which enables quick examination of your bandwidth performance.

The dashboard's simplicity of use and accessibility of custom reports make it a popular option for clearly and swiftly examining network performance information.

Solarwinds NetFlow Traffic Analyzer provides the following capabilities:

  • By drilling down into any network element, you may analyze network traffic patterns across months, days, or minutes.
  • Monitors Cisco NetFlow, sFlow, Juniper J-Flow, Huawei NetStream, and IPFIX flow data and detects the applications and protocols that use the most bandwidth.
  • With NetFlow analyzer insights, you can do faster troubleshooting, boost productivity, and get more visibility into malicious or corrupt network flows.
  • Be able to intervene immediately if an unexpected shift in application traffic occurs. Additionally, you may configure alerts to be informed when a device ceases to deliver flow data, allowing you to quickly resolve the issue.

Solarwinds NTA

Figure 3. Solarwinds NTA

Is QoS Used in Next-Generation Firewalls?

Definitely, yes. You can consider a next-generation firewall as a solution that combines a packet-filtering firewall and quality of service (QoS) functionality. They may act as content filters and offer QoS features, ensuring that higher-priority apps get more bandwidth. Network quality of service, or QoS, includes bandwidth control and prioritizing traffic. Traffic prioritization and bandwidth management may be used together or separately depending on the sort of traffic you're dealing with.

The basic objective of QoS in the next-generation firewall is to apply rate-limiting to chosen network traffic, whether it is individual or VPN tunnel traffic, to ensure that all traffic receives an equitable part of the available bandwidth. A flow may be characterized in a variety of different ways. QoS may be applied to a mix of source and destination port numbers, source and destination IP addresses, and the IP header's Type of Service (ToS) byte in the security appliance.

The NGFW is capable of reading and writing type of service (ToS) field markers from DiffServ Code Point (DSCP). The markers enable the NGFW to be informed of the network equipment's priorities. The markers enable you to combine the firewall with other network equipment that manages QoS in your network or the network of your ISP.

There are three kinds of QoS you can implement on the NGFW: Policing, Shaping, and Priority Queueing.

  • Policing: Traffic exceeding a defined limit is halted by police. Policing is a technique for ensuring that no traffic exceeds the configured maximum rate (in bits/second), ensuring that no one traffic flow or class may take over the whole resource.
  • Shaping: Traffic shaping is used to align device and connection speeds, therefore reducing packet loss, variable latency, and link saturation, all of which may result in jitter and delay.
  • Queueing: Priority queuing enables you to put a certain kind of traffic in the Low Latency Queue (LLQ), which is handled ahead of the ordinary queue.

What is the Difference Between Network Slicing and QoS?

Network slicing is a subset of virtualization that enables the operation of numerous logical networks on top of a common physical network infrastructure. The primary advantage of network slicing is that it creates an end-to-end virtual network that encompasses not just networking but also computation and storage services. The goal is to enable a physical mobile network operator to split its network resources in order to allow for multiplexing of very distinct users, referred to as tenants, on a single physical infrastructure.

Network slicing is distinct from QoS in that it enables the creation of end-to-end virtual networks that span computing, storage, and networking operations. Existing QoS techniques are all point solutions that provide just a piece of network slicing capability. VoIP traffic may be distinguished from other forms of traffic such as HD video and web surfing using QoS. However, QoS cannot differentiate and handle differently the same kind of traffic (e.g. VoIP traffic) originating from multiple sources, but the same type of traffic may be discriminated against and treated differently when allocated to distinct slices.

Additionally, QoS lacks the capacity to execute any kind of traffic isolation. For instance, IoT traffic from a health monitoring network (e.g., linking hospitals and outpatients) is often subject to stringent privacy and security regulations, including the location of data storage and who has access to it. This cannot be performed using QoS because it lacks functionality for managing the network's computation and storage resources. All of QoS's reported shortcomings will be addressed by the network slicing technologies now being developed.

What is the Difference Between QoS vs. HQoS?

In the Differentiated Services (DiffServ) architecture, Hierarchical Quality of Service (HQoS) ensures the bandwidth of various services for customers via the use of a hierarchical queue scheduling mechanism. Traditionally, QoS is used to schedule traffic on a single level. A single port can distinguish between services but not between users. HQoS employs hierarchical scheduling based on many layers of queues and separates traffic across users and services to provide a finer quality of service assurance.

HQoS is not a stand-alone QoS solution, as is the case with traditional QoS based on the DiffServ approach. It enhances traffic management by introducing hierarchical scheduling based on standard QoS.

In the QoS DiffServ approach, packets are assigned to various queues depending on their priority mappings to the device's internal priorities. According to the scheduling algorithm (such as priority-based SP scheduling or weight-based DRR/WRR/WDRR/WFQ scheduling), the scheduler decides the order of transmitting packets in separate queues.

scheduling models of traditional QoS and HQoS

Figure 4. Scheduling models of traditional QoS and HQoS

The following sections discuss the distinctions between standard QoS and HQoS via the use of two cases.

  • Traditional QoS: One-Level Scheduling Mechanism: Assume that data, voice, and video service packets enter distinct queues. These queues are divided into high-priority queues, medium-priority queues, and low-priority queues. They all employ the SP scheduling algorithm. When packets are scheduled to exit a queue, they are planned according to their priority. The first packet is a voice packet, followed by a data packet, and finally a video packet. One-level scheduling in Traditional QoS

Figure 5. One-level scheduling in Traditional QoS

Conventional QoS organizes traffic on a port-by-port basis, with a single port able to discern service priority but not user preferences. As a result, several forms of traffic with the same priority but originating from distinct users are consolidated into a single port queue (PQ), where they compete for resource allocation. As a result, differentiated services cannot be provided to distinct users.

  • HQoS: Hierarchical Scheduling Mechanism (Two-Level Scheduling): Assume there are both ordinary and VIP users. Both sorts of users have access to data, audio, and video. HQoS supports two scheduling levels: user-based and service-based scheduling. User-based scheduling guarantees that VIP users' packets are prioritized for transmission. Service-based scheduling is used to prioritize critical services for each user, similar to how one-level scheduling is used in traditional QoS. Two-level scheduling of HQoS

Figure 6. Two-level scheduling of HQoS

Accompanying the implementation of HQoS in a port, traffic management may be abstracted, as seen in the following image. After allocating user bandwidth to distinct users based on port bandwidth, service bandwidth is assigned to distinct services depending on each user. HQoS traffic management

Figure 7. HQoS traffic management

Through hierarchical scheduling, HQoS distinguishes both services and users, ensuring that VIP users' and high-priority services are processed preferentially, and assigns guaranteed bandwidth to distinct services and users, implementing more precise traffic management than standard QoS.

QoS History

Although the necessity for differentiated network performance for different types of traffic is widely understood, this requirement evolved as remote access to computer applications became more significant and the economics of providing converged networks became apparent.

Numerous attempts at layer 2 technologies that tag data with QoS attributes have found traction in the past. Frame relay, asynchronous transfer mode (ATM), and multiprotocol label switching (MPLS) are all examples (a technique between layers 2 and 3).

In 1984, X.25 guideline established the concept of quality of service in the form of a throughput class, which is defined as the maximum number of bits per second that can be transported across a given Virtual Circuit (VC). The default throughput class was determined using the access link's speed and applies to all logical channels on the link. This is configured at the time of subscription and is applied until overridden during call setup.

Frame Relay eliminates the flow control methods found in X.25, leaving the network susceptible to congestion, which can result in packet discard. When a network node detects congestion in one direction, it tags traffic flowing in that direction with the FECN (Forward Explicit Congestion Notification) bit in the frame header, whereas traffic traveling in the other direction is tagged with the BECN (Backward Explicit Congestion Notification) bit. Recovery is outsourced to higher-level protocols (usually TCP).

ATM standards define a variety of distinct traffic classes, the most frequent of which is variable bit rate (VBR). Here, quality of service (QoS) is defined in terms of accuracy, dependability, and speed. When a PVC is created, the network analyzes whether adequate resources are available to support the desired performance without jeopardizing current virtual circuits' contracted levels of service.

There are two major architectures for achieving QoS: the integrated services model, or IntServ, and the differentiated services model, or DiffServ. IntServ is the older of the two protocols and requires capacity reservation across all network nodes to ensure the appropriate level of service can be given. Typically, the Resource Reservation Protocol (RSVP) (Braden et al., 1997) is used to reserve enough resources to support the specified load traffic. Traffic classification was handled on a per-flow basis, which created scalability challenges and limited widespread implementation; nonetheless, the core principles have been incorporated into MPLS traffic engineering approaches.

The scalability hurdle was overcome by integrated services architecture due to the same issue that afflicted X.25 and older IBM networking: you simply cannot offer individual QoS guarantees to millions of flows transiting the same high-speed link. As a result, all designs for high-speed service providers have differentiated services architecture.

DiffServ's method is based on the Per Hop Behavior (PHB) of specific traffic categories rather than individual flows, with traffic categorized according to the TOS byte in the IP Protocol header. The arrangement of this byte was described in RFC 791, modified in RFC 1349, and re-specified in RFC 2474 to use the six most important bits of the TOS header to establish the Differentiated Services Code Point (DSCP), which is used to categorize data on a per-hop basis.

Despite the fact that these network technologies are still in use today, this type of network waned in popularity following the emergence of Ethernet networks. Ethernet is by far the most widely used layer 2 technology today. On a best-effort basis, conventional Internet routers and network switches operate. This equipment is less expensive, simpler, and faster than earlier, more complicated technologies that provide QoS mechanisms, and hence more popular. Ethernet can optionally employ 802.1p to indicate a frame's priority. Originally, each IP packet header contained four types of service bits and three precedence bits, but they were not widely respected. Later on, these bits were renamed Differentiated services code points (DSCP). With the emergence of IPTV and IP telephony, end users are increasingly able to use QoS algorithms.