A network service is a capability that facilitates a network operation. This is typically provided by a server, which can run one or more services and can be based on network protocols running at the application layer in the open systems interconnection (OSI). Some examples include the domain name system (DNS), voice over internet protocol (VoIP), and dynamic host configuration protocol (DHCP).
Over time, enterprises have shifted thinking that their networks were once just a tactical necessity, but are now a strategic asset to a core of business strategies, especially in the case of sharing data internally, distributing data externally, and global collaboration. A 2011 survey found that only 38 percent of participating IT leaders perceived their network as a strategic asset; that number grew to forty nine percent in 2015.
Networks can trace their use back to the US defense department projects in the 1960s, out of which the internet was born. The early days' evolution was as a military-only network, where the military could have a decentralized, indestructible communication network. The desire for a distributed communication model, which could continue to work in worst case scenarios (such as nuclear strikes), led to DARPA's creation of ARPANET, one of the earliest iterations of computers talking to each other.
Comparison of a centralized versus distributed network.
Indestructibility was key to the early project, and with a wish to avoid the hub-and-spoke model used by telephone networks of the time, where an operator (the hub) would patch two people (the spokes) to each other. The RAND corporation developed a technology in the early 1970s, later called packet switching, that allowed users to send secure voice messages through an entire network, or web, of carrier lines, without need to travel through a central hub. In this way, there was not a central hub that could impair the network if destroyed.
ARPANET was eventually expanded to military installations, third-party contractors, and a handful of universities in the United States. These early systems used interface message processors, which were designed to organize and receive data coming in and out of the network and were the earliest versions of modern routers. ARPANET relied, as well, on leased telephone lines, similar to the subsequent commercial internet. And in the 1970s, ARPANET was connected to NORSAR, a US-Norwegian system designed to monitor seismic activity from earthquakes or nuclear blasts, over satellite. NORSAR then connected to computers in London and, eventually, in other parts of Europe.
The early institutions that connected to ARPANET, including UCLA, Stanford, UC Santa Barbara, and the University of Utah, were considered hosts of the network. These hosts were a necessary part of the network as a physical link that is connected to the Internet and directs traffic by routing packets of data to and from other computers connected to it. In a normal network, a specific computer is connected to the internet through a host, which is identified by an Internet protocol (IP) address; the IP address, in turn, serves as a gateway for multiple computers while also referring to a single location on the Internet.
Through the late 1970s and early 1980s, similar network ideas to ARPANET emerged, including the ALOHA system built at the University of Hawaii, which was a packet-switched network operated over radio, rather than leased telephone lines. This led to PRNET, which was another packet-switching radio network in the San Francisco area. This led to ARPANET, PRNET, NORSAR, and SATNET (which was established to monitor Soviet nuclear testing via satellite); these used packet-switching in different ways and with the next step to interwork the networks so they could function as an apparent whole. This led to the adoption of the Transmission Control Protocol and Internet Protocol, and they became the cornerstone of the new network of networks. The approach meant any network capable of "speaking" TCP or IP was free to join the internet. There were no gatekeepers and no administrators.
As the networks grew and became more complicated, email, or electronic messages, were able to be delivered to an individual's personal folder, similar to leaving a note on a person's desk. With the increased complexity, there was a need for a more sophisticated system. Computer programmer Ray Tomlinson would be credited with inventing the naming system used, with the @ symbol representing the server (or host), which meant using a name@host.com would tell the host ("host.com") to leave a message in the folder belonging to "name". The use of email continued to grow with further commercial developments; for example America Online made connecting to email easier than at its inception. Internet service providers began packaging email accounts with internet access and almost all web browsers included a form of email service. The further development of free email services such as Hotmail and Yahoo! Mail increased adoption of the service.
The internet protocols and technologies—which included the Transmission Control Protocol (TCP and Internet Protocol (IP) suite of protocols—began to be standardized in the late 1980s and early 1990s as access was extended beyond institutions and towards regular people. Connecting over the internet at this time was without the World Wide Web, which made information difficult to search for, and the text-based system was considered dense and uninviting.
Tim Berners-Lee eventually developed the concept of the World Wide Web, a decentralized repository of information that was linked together and shareable with anyone connected to it. The first webpage was built by Berners-Lee in 1993, and the software was opened up to the public domain. Berners-Lee also created the first web browsers, named Nexus, but it would be the creation of Mosaic and Netscape Navigator that made the internet easier to navigate. With the introduction in 1996 of the 56k modem, there was an explosion of internet service providers and users connected online. This new commercial network, built upon the earlier network technologies, continued to work over loaned phone lines that the network would call over.
There was a need for solutions to problems of the early dial-up internet that needed to be solved before it could be made more available. These problems included a need to allow for interactivity between browsers and servers, a facilitation of personalization of web content, and to overcome the insecurity of httpprotocol. The last was solved by the introduction of HTTPS, an encrypted version of the httpprotocol. Browsers evolved capabilities such as "cookies" and specialized plug-ins to allow them to handle audio and video and other files, and eventually JavaScript turned web pages into small virtual machines.
As well, dial-up gave way to broadband, and in 2004 more people had access to the new broadband compared to the original dial-up. Broadband works differently than dial-up, with no need to call out over the phone line and the modems stay connected unless turned off. Coupled with the development of Wi-Fi, internet continued to become more available, more useful, and offered a faster and less frustrating experience. It also became more mobile, with technologies such as the Palm Pilot and the Apple iBook laptop in 1999, which included a Wi-Fi card.
These technologies have since given way to cellular data, or mobile broadband, which allows a user to connect to the internet through a cell phone. At the end of 2013, there were about 1.3 billion smartphone subscriptions globally, while at the end of 2018 those increased to around 5.3 billion. The first truly mobile data standard, developed after the Wireless Application Protocol (WAP) and often considered dial-up for phones, was released in 2003 and known as 3G. This technology allowed for more calls and texts to be sent. The 3G technology has since given way to 4G and 5G technologies, and the smartphone's computing power has increased with the increased use of smartphones. And the use of the internet has gone from large connected computers to smaller, mobile computers and devices.
Considered a key technological development in both the use of the Internet and the development of the internet, there has been consistent improvement and development of Wi-Fi technologies. One such example is Velmenni, an Estonia-based business, which is developing a technology called Li-Fi, intended to deliver faster and more secure connections to users. Li-Fi operates through LED bulbs connected to a computer or laptop and has delivered a consistent speed of 1Gbps in real-world testing, while lab testing has reached 224Gbps, which would be fast enough to download a dozen high-quality films in a second.
Furthermore, the development of Li-Fi decreases the amount of hardware used by Wi-Fi by using a microchip controller, increases the security of Wi-Fi by controlling and communicating over light rather than radio waves, and streamlines wireless networks by integrating them with solar energy.
Meanwhile, Qualcomm is developing wireless technology through the company's small cells project, which focuses on hyper-dense deployment of cells that act as a Wi-Fi hotspot but are the size of a wireless access point. China Mobile is developing a network technology that is software-based and similar to the principles of cloud-computing.
Although not originally called 1G, the first generation of mobile network emerged in Japan in 1979, before being adopted by other countries (such as the US in 1980 and the UK in 1985). Based on a technology known as advanced mobile phone system (AMPS), which uses frequency division multiple access (FDMA) modulation, 1G networks offered a channel capacity of 30KHz and a speed of 2.4Kbps. 1G networks allowed voice calls, but suffered from reliability and signal interference issues and had limited protection from hackers.
The first generation of mobile networks was not superseded until 1991, when 2G networks were introduced. These networks were based on digital signaling technology, known as Global System for Mobile Communication (GSM), which increased security and capacity. The 2G network offered bandwidths of 30KHz to 200KHz and allowed users to send SMS and MMS messages at low speeds and up to 64Kbps. Improvements to GSM technology led to the so-called 2.5G, which incorporated packet switching in the form of GPRS and EDGE technology. The 2.5G network offered data-rates up to 144Kbps and enabled users to send and receive emails and browse the internet.
Third generation (3G) networks superseded in 1991, and were known as UMTS in Europe and CDMA2000 in the United States. The new network technology offered a change in the way mobile phones were used and viewed by users. These devices were used less for voice calls and were increasingly used for social connectivity. Also based on GSM, the aim of 3G was to support high-speed data. The original 3G technology allowed rates up to 14Mbps. This offered the ability to transmit greater amounts of data at higher speeds and allowed users to make video calls, access the internet, share files, play online games, and watch television online. Where 2G allowed a 3-minute MP3 song to be downloaded in around six to nine minutes, the same file would take anywhere between eleven to ninety seconds to download on a 3G network.
The introduction of 4G in 2009 allowed for the development and adoption of smartphones and hand-held mobile devices. This was the first generation to use Long-Term Evolution (LTE) technology that delivers theoretical download speeds between 10Mbps and 1Gbps, offering end users better latency, improved voice quality, instant messaging services, and social media. This includes better quality streaming and faster download speeds. 4G also offered the first IP-based mobile network that handled voice as another service and the technology to accommodate the Quality of Service (QoS) and rate requirements required by applications including wireless broadband access, multimedia messaging service (MMS), video chat, mobile TV, HDTV content, and Digital Video Broadcasting (DVB).
These networks were, in short time, struggling to to cope with the demands placed on them, driven by emerging technologies such as Augmented Reality, autonomous vehicles, and the Internet of Things (IoT). Further, these and other emerging technologies tend to be bandwidth intensive and require high speeds at lower latencies than previous networks were capable of providing. This led to the formation of the International Telecommunications Union (ITU) in 2015, to define the requirement specifications for the next generation of mobile networks to be developed before the 4G networks would reach capacity.
The specification for 5G, contained in the document ITU-R IMT-2020 (5G), developed by ITU, represented a step-change in performance over 4G and aimed to address the requirements of the emerging applications. Throughputs up to 10Gbps work to satisfy the growing need for bandwidth, latencies of 1mSec to enable near-real time response rates, and connection densities of 1000 devices per square kilometer in order to support the growing number of IoT devices and sensors.
The deployment of all of the 5G capabilities defined in IMT-2020G requires new networks, which in turn require significant investments by operators and time in order to enable a full roll-out of 5G networks. To ease the migration path and allow operators to get to market early, 3GPP defined 5G Non-Stand Alone (NSA) technology. 5G NSA enables 5G services to be provided using existing LTE infrastructure with the throughput of existing macro cells able of being increased by adding extra MIMO layers. These layers allow operators to use the existing spectrum in the MIMO sweet spot (around 3.5GHz) to offer consumers 5G networks and services to new consumer 5G handsets.
Further, the release specification for 5G NR Stand-Alone (SA) technology, scheduled for completion in early 2020, addresses the specifications for mmWave technology based on spectrum allocation decisions following the ITU's World Radio Conference in October 2019.
With the evolution of networks and network technology, there has been a need for carriers (or telecommunication companies) to change the way they have delivered managed network services, although this evolution has been slow and has not moved far past the core architecture of most network services, the hub-and-spoke system. In this architecture, branches (or spokes) talk to the data center (a hub) over a managed network with a separate firewall in the middle. This type of WAN does not support the shift to cloud computing, nor does it easily integrate the increase in mobile users that require network access from anywhere.
These managed networks work for branches and physical locations, which leaves mobile workers outside of the network. Alternative methods of connectivity, such as VPNs, allow a mobile worker to access these closed networks.
One of the first real changes in evolving managed network services market was the network function virtualization (NFV). This included the virtualization of all the different network functions, including firewalls, various orchestration solutions, and VPN solutions. The virtualization does not change the core dynamic of the network and was not seen as enough in order to change the network in order to achieve network agility and flexibility.
This further saw a switch from physical routers and switches to virtual routers and virtual switches. And, in turn, a development towards software defined networking (SDN) aimed to simplify network management with the continued evolution of applications. While these applications have become more loosely coupled and fine grained, the network has remained course grained and strongly coupled with many definition levels.
There followed a further move towards the AWS model of managed network services. This sees a managed network and managed servers, managed storage, and other cloud-based computing solutions that other network service operators do not offer, such as telecommunication companies. There came an increase in managed network service providers, such as Cat Networks, Microsoft, Aryaka, and Meta Networks. These are cloud-native services and everything resides in the cloud with customers able to subscribe to a service. The provider establishes a private global network comprised of numerous points of presence over a multi-carrier Tier 1 backbone. The service provider controls the routing latency of packets on a global scale over a predictable and SLA-backed backbone. And, using multiple links and load-balancing them, the service provider can offer reliability, high availability, and consistent performance across an organization.
With the rise of cloud native application developers, there is a preference to be concerned with secure intranet connectivity and manufacturing partner networks over concerns with what is happening in a network, such as subnets or routes. The developers do not want to focus on network details, but prefer to focus on the application level and leverage multi-cloud applications where possible. An example of multi-cloud cloud native connectivity is the Network Service Mesh (NSM), which allows individual workloads to connect to multi-cloud network services in order to meet their needs, independent of infrastructure.
The managed network providers can connect their data centers, branches, and mobile users to the global network. The network also peers with public clouds and software-as-a-service applications, and offer customers direct and secure access. As well, security services, such as firewalls, anti-virus, and anti-malware, and IDS or IPS services can be integrated into the network and are available anywhere with the aim of the architecture to solve problems that legacy WAN architecture cannot.
With the COVID-19 pandemic of 2020, and many countries introducing stay-at-home orders that led to employees working from home, the focus on networks and the role of digital infrastructure in the functioning of contemporary business and society was brought into relief. The important features of digital infrastructure is the ability to bridge distances and make it easer to meet needs in terms of resource utilization, collaboration, competence transfer, status verification, privacy protection, security, and safety. This remains true as the transition to 5G connections begins. These managed network services are necessary for workers to be able to connect to networks and for these networks to secure access and services for workers. This has meant an increased move away from users remaining on managed hub-and-spoke networks towards more distributed networks that can be accessed from anywhere.
Part of the overall evolution of network services includes the use of cellular networks and the ubiquity of cellular coverage, which have allowed for sufficient connectivity for IoT applications. Some of these services have high demands on low latency, such as remote driving of automated commercial vehicles, which current 2G to 4G networks do not always provide. However, 5G offers improved network capacity, higher data rates, lower latency, and improved experience and reliability in different use cases, such as connected vehicles.
Different network needs for different industries as they go digital offers an opportunity for communication service providers (CSP) to provide these services, so long as a broader set of network capabilities are included, such as:
- Network slicing to realize specific and dedicated end-to-end services over CSP networks, allowing for the building of logical networks on top of a common and shared infrastructure
- Zero trust architecture, including an evolving set of cybersecurity paradigms that move defense from static, network-based perimeters to focus on users, assets, and resources
- A need to expose network capabilities as CSPs to explore the specific needs of enterprises
- A capability for the network to evolve to a data and intent driven cognitive network with a higher level of automation and optimization
The development of digital infrastructure and networks is the ability of these services to bridge distances and make it easier to efficiently meet societal needs in terms of resource utilization, collaboration, competence transfer, status verification, privacy protection, and security and safety. This infrastructure and network capability can, in turn, support other industries in the development and deliverance of digital products and services, such as health care, education, finance, commerce, governance, and agriculture.
The concept of networks continues to evolve towards a vision of a future network capable of delivering limitless performance to satisfy the needs of humans, things, and machines by enhancing multidimensional coverage, stellar capacity, and augmenting capabilities. And as distributed compute and storage continues to evolve, the lines between device, network, and cloud become increasingly blurred until everything can be viewed as a single, unified, integrated execution environment for distributed applications. In a network compute fabric, connectivity, compute and storage would be integrated and interact to provide maximum performance, reliability, low jitter, and millisecond latencies for the applications they serve.
