Early
on, the Internet was believed to offer a single level of service, wherein all
data packets traversing the network were treated with equity. Today, however,
some areas of the Internet have been seen to exhibit high levels of congestion
and consequently poor quality, while other areas display consistent levels of
high quality service. It thus became evident that the Internet may not be offering
a single level of service quality after all / anymore. Customers are now beginning
to insist more clearly on specifications and definitions for a consistent
service quality that they wish to be provided, even as network service
providers are seeking ways of meeting such requirements and specifications.
In
light of this, modern QoS implementations now reflect an array of various
methodologies and technologies that are being applied in order to manage
traffic in such a way that reduces packet loss, latency, and jitter as content
(which could be voice, video and delay-sensitive data) travels through the
network. This sometimes also demands that limited communications resources
(such as bandwidth, throughput, etc.) are effectively appropriated in order to
ensure that the priority specifications that have been set for different types
of content traversing the network are adhered to. Hence, when new applications
and services emerge on the network, they are required to provide services that
can be quantified and measured in terms of the QoS implementations that have
been deployed on the network. For businesses and organizations that provide
internet services, this requirement serves to ensure that the services that
they offer do meet the network traffic requirements of sensitive corporate applications,
while also preventing the degradation of quality, which can be caused by packet
loss, delay, and jitter. Service-level Agreements (SLA) between such service
providers and their clients help to guarantee a certain agreed (minimum) level
of performance quality that provisioned services must meet.
Quantifying Quality of Service
Several
considerations that attempt to quantify or measure QoS often do so in light of
the following parameters, among others:
- Jitters: could result from congestion, drifts in timing specifications, or the re-routing of content in transmission. This is able to degrade the quality of voice and video communication over the network.
- Latency: the time it takes a packet to travel from source to destination is known as latency. In ideal implementations, latency should be as close to zero as possible; or close enough to be negligible in best case implementation. When latency is high, echoes and overlaps could appear in voice communications, as well as delays in video streams.
- Packet Loss: which is often the result of network links becoming congested in such a way that routers and switches begin to drop packets in transmission. This could translate to jitters, high latency / delays, and gaps in real-time applications for voice and video communications.
- Bandwidth: is basically the capacity of a network transmission link to convey a certain amount of content from source to destination within a certain period of time. By reserving, reallocation and managing bandwidth, while also prioritizing applications that require more resources than others, many QoS implementations are able to optimize the quality of network services.
- Mean Opinion Score (MOS): is an aggregated metric that rates the quality of voice transmissions using a five-point scale; where an MOS score of “5” signifies the best quality, while a score of “1” signifies the lowest quality.
Approaches to Implementing QoS
Three
widely-applied models have been documented for implementing quality of service
in modern networks. They are:
- Best Effort. Here, all packets receive equal priority in transmission, with no guarantees of delivery. This approach is often featured in networks that either have no QoS policies configured, or contain infrastructure that do not support QoS.
- Integrated Services (IntServ). This approach features a reservation of bandwidth along specific network transmission paths, which could be appropriated to requesting applications. Implementing this approach requires dedicated network hardware and protocols, and is also limited in scalability due to its resource intensiveness.
- Differentiated Services (DiffServ). Here, network devices (routers and switches) are configured to service multiple classes of traffic that have been prioritized differently. Every network traffic must belong to one of the configured classes, in order to be serviced. By applying techniques for queuing and (re-)prioritization, the QoS of the DiffServ model can be optimized.
Furthermore,
Multiprotocol Label Switching (MPLS) networks provide dedicated links that
provision end-to-end quality network services to organizations via a single
transmission path. SLAs for MPLS networks include QoS specifications for bandwidth,
latency, and uptime; however, few organizations are able to cope with the
associated high costs.
Similarly,
Software-defined WAN (SD-WAN) deploys different modes of connectivity that
could include MPLS and broadband. By monitoring the state of current network
connections for performance issues, SD-WANs leverage multiple connectivity modes
to fail-over the states of connections that continue to perform poorly, and
then look for alternative, better-performing connection(s).
QoS Tools and Mechanisms
Some
tools and mechanisms exist, which aim to manage the quality of data traffic,
and also ensuring to maintain the QoS requirements that have been specified in SLAs.
Depending on their management functions within the network, these QoS tools and
mechanisms could be categorized as:
- Classification & Marking – these are tools and mechanisms that help to differentiate between applications by sorting and grouping packets into different traffic types as they traverse the network routers, switches, and access points. By marking / labelling each packet as a member of a certain network class, network devices are able to recognize and handle the packet as belonging to such class.
- Congestion Management – these tools rely on the classification and marking of packets, in order to decide what transmission queue that such is to be placed in. Considerations such as priority queuing; first-in, first-out; and, low-latency queuing typically feature in these operations.
- Congestion Avoidance – tools and mechanisms such as: weighted random early detection and random early detection, monitor network traffic for congestion and will drop low-priority packets when congestion occurs.
- Shaping – tools and mechanisms include buffers, Generic Traffic Shaping and Frame Relay Traffic Shaping. They manipulate the traffic entering the network, and prioritize real-time applications over those that are not time-sensitive; such as email and messaging. Similar to shaping, traffic policing tools focus on throttling excess traffic and dropping packets.
- Link efficiency tools maximize bandwidth use and reduce delay for packets accessing the network. Approaches could include Real-Time Transport Protocol header compression, Transmission Control Protocol header compression, and link compression.
The Role of Government
Government
plays a crucial and indispensable role in ensuring the quality of Internet
services that are provisioned to users and businesses within their
jurisdictions. Through their regulatory bodies, they have the responsibility to
ensure that good quality Internet service is attained and maintained by service
providers that operate within the jurisdiction.
In
the pursuit of this responsibility, Governments are encouraged to adopt a
consumer-oriented approach that focuses on delivery of quality services at
reasonable and affordable costs; as well as facilitate the introduction of new
modern services, and the expansion of existing services into modern and
innovative quality services delivered at reasonable and affordable costs. Thus,
Governments:
- Set QoS standard, and specify the QoS parameters against which performance shall be evaluated, with the associated target levels. Governments may periodically amend these parameters or the targets to align with evolving trends and development.
- Specify methodologies for measuring performance against each Quality of Service parameter, and would periodically carry out measurement assessments against QoS parameters as deemed necessary.
- Receive reports and carry out occasional audits / assessments of the QoS provisioned by licensees within their jurisdictions; and could apply appropriate sanctions in cases of non-compliance by licensees.
- Handle complaints from consumers seeking redress due to dissatisfaction regarding the services rendered to them by service providers. Such complaints may relate to arbitrary disconnection, poor services delivery, supply of sub-standard equipment, and delayed restoration of service, poor picture quality, etc.
- Publish results of QoS measurements and assessments carried out by regulatory bodies to periodically evaluate the QoS performance of licensees operating and provisioning services within the jurisdiction.
In
conclusion, discussions are still ongoing pertaining whether or not it is
possible to achieve ubiquitous end-to-end QoS on the Internet; particularly
considering the fact that the decentralized nature of the Internet does not essentially
allow for the traffic to be differentiated using exact or widely similar
mechanisms. The polarization of such discussions have polluted expectations
regarding QoS, and how QoS technologies might be realistically deployed on a
global scale.
These notwithstanding, it would be chaotic to discuss the modern
trends and realities associated with the quality of Internet service in certain
jurisdictions with the role “government” removed from the picture. The role of
Governments is the adhesive that binds together the contemporary discussions
relating to Quality of Service as an Internet Governance discourse. As the
global landscape gradually gravitates towards more inclusive, participatory,
and collaborative approaches to Internet Governance which leaves no stakeholder
behind, it is hoped that citizens would be able to insist rightly and
knowledgeably on quality Internet service; even as governments become proactive
and responsive to the duties and mandates of their roles and responsibilities
in ensuring good quality Internet service within their jurisdictions.
Bibliography
[1] Huston, G.
(2012, September). The Concept of Quality
of Service in the Internet. APNIC. Retrieved from https://ripe65.ripe.net/presentations/67-2012-09-25-qos.pdf
[2] Kurbalija, J.
(2014). An Introduction to Internet
Governance (6th edition). DiploFoundation. Retrieved from https://www.diplomacy.edu/sites/default/files/An%20Introduction%20to%20IG_6th%20edition.pdf
[3] Zhao, W. (2000,
April 20). Internet Quality of Service.
Retrieved from http://www1.cs.columbia.edu/~zwb/my/talk/InternetQoS.pdf
Image Source: https://encrypted-tbn0.gstatic.com/images?q=tbn%3AANd9GcQ99NEEcC7UCh3lbx_ZcTKH4GihFdGygICkKDCckzA7UOpUi1oA
[Retrieved on November 20, 2019]
Author Biography
 |
Innocent Adriko is an Information Technology Scientist from the
Uganda Institute of ICT. He holds a Diploma in Information Technology
Science, with great interest in Internet Policy and Governance, and Youth
inclusion in Internet Governance. He has undergone a number of trainings
especially on Internet Governance, Youth Peacebuilding and Human Rights. He
is a member of the ICANN community which he joined as a Next Generation at ICANN65
Policy Forum; and is also a member of the Internet Society (ISOC) and one of
the ISOC 2019 IGF Youth Ambassadors. Innocent is most passionate about
International Relations and Diplomacy with great focus on global matters of
concern such as Internet Governance, Peace and Security and Climate action.
Country: Uganda
|
No comments:
Post a Comment