第五次网络浪潮
为了成功地满足未来苛刻的需求,网络的第五波将不仅仅是无线电的渐进式改进。
为了成功地满足未来苛刻的需求,网络的第五波将不仅仅是无线电的渐进式改进。
Growing video consumption and new applications will continue to drive exponential data traffic growth on mobile networks; video content itself is getting heavier with 4K video. By 2021, mobile video data consumption is set to increase by 870% from 2016.Further AR\/VR devices will have a massive multiplier effect on the total data traffic volume and will put much higher \u2018throughput\u2019 demands on the network of the future. Massive IoT will disproportionately push the network limits on total number of \u2018connections\u2019 as the industry embarks to connect 20B devices and more.And new applications like driverless cars and self-flying drones will demand much lower network \u2018latency\u2019. Network of the future has to meet these unprecedented demands on: throughput, connections and latency.<\/p>
Some of you might recall the massive challenges CSP networks faced in the early days when smartphones were launched. Network engineers on some days were dealing with \u2018data tsunamis\u2019 as smartphones pushed the network\u2019s limits on data throughput. And on others days the same network engineers would be dealing with \u2018signalling storms\u2019, when un-optimized apps were creating havocs on network\u2019s signalling traffic. It is worth noting that networks experienced all these challenges with just one type of \u2018thing\u2019: smartphones. We can foresee what\u2019s to about come with IoT, when networks will have to deal with massively varying use-cases from massively varying devices. Massive IoT further exacerbates security threats, last year\u2019s Mirai DDoS botnet attack clearly exemplifies the threats network of the future will face. It will no longer be viable to profitably operate and secure the network of future with traditional tools and processes. Hyper-automation is the only way forward to meet the hyper-demands of IoT. Big Data, AI, Machine Learning, Analytics will all play a key role in driving hyper-automation of the network of future.<\/p>
Furthermore, for service providers to seize the emerging opportunities networks will have to deliver more-for-less. According to GSMA, from 2016 \u2013 2020 data traffic on mobile networks is projected to grow at 47% CAGR, while revenue is projected to grow just at 2.1% CAGR during the same period. The math couldn\u2019t be simpler; to keep their networks sustainably profitable, service providers have to find solutions that aggressively bend their cost curve. To achieve this, network of the future requires: new architecture. Traditional network appliance models no longer profitably scales to meet the future needs, linear-scale must give way to hyper-scale. Service providers need an architecture that enable networks to be elastic, programmable and agile. Software Defined Networking (SDN) and Network Functions Virtualization (NFV) are the forces that are shaping up the architecture for network of the future. Web-scalers have already proven building hyper-scale software-defined infrastructure leveraging COTS platform and merchant silicon. SDN\/NFV promise to do just that for operators; driving optimal resource utilization and enabling them to innovate their networks at the speed-of-cloud.<\/p>
We strongly believe the three key pillars of the 5th wave of networks are going to be:<\/p>
Spectrum is the lifeblood of a mobile network. To dramatically increase network throughput and significantly lower latency, network of the future requires new access. 5G NR is shaping up to address these requirements by significantly improving spectral efficiency and by leveraging higher frequency bands (>6GHz), including mmWave, where much higher channel bandwidths are available. Network of the future will leverage both licensed and unlicensed spectrum, as well as new shared spectrum options like: CBRS 3.5GHz in the U.S. Numerous 5G NR trials are already underway as service providers mature RF technology for 28GHz, 31GHz and even 60GHz mmWave. We are also witnessing dis-aggregation of the base station itself with vRAN (virtualized RAN). vRAN will dramatically lower the cost of the RAN equipment. RRH (Remote Radio Heads) are being decoupled from BBU (Baseband Unit), and BBU itself is being virtualized. And that brings me to NFV.<\/p>
<\/p>
SDN and NFV are laying the foundation of the network of the future by adapting the Cloud technologies for the carrier networks. We are already witnessing the growing importance open-source-software will play in the network of the future. The industry is humming with numerous open source projects including: OpenStack, OpenDayLight, ONOS, ONAP, OSM, CORD and more. The software architecture itself is evolving to microservices, as service providers demand solutions to be cloud native. Old stovepipe OSS stacks are starting to give way to new-age Management and Orchestration (MANO) solutions that will bring unprecedented agility, enabling service provider to quickly design, realize and offer new services. New architecture will enable service providers to achieve closed-loop-control, such that the network elastically scales up and down depending on varying traffic conditions. Finally, the new architecture will enable network slicing, enabling service providers to programmatically create network slices catering to varying demands of various IoT use cases.<\/p>
<\/p>
Pairing vast amounts of network data collected in data lakes with the vast amount of compute power now creates the potential to unleash the power of machine learning and AI algorithms to achieve unprecedented levels of hyper-automation. Future networks will be self-aware and self-optimizing. Leveraging predictive analytics these networks will achieve optimal asset utilization and near real-time auto-scaling. They will be able to predict faults and self-heal. Hyper-automation will pave the way for self-defending networks, that will be able to detect traffic anomalies and proactively prevent against security attacks like the recent Mirai massive DDoS Botnet attack.<\/p>
<\/p>
To conclude, the future of the communication industry looks very bright as the demand for data continues to grow. IoT presents unprecedented growth opportunity for service providers. To successfully seize these growth opportunities, the service providers must transform their networks to achieve unprecedented levels of agility, elasticity and automation. The 5th wave of networks will be go much beyond the Radio innovation. New Radio (NR), New Architecture (SDN\/NFV) and Hyper Automation (AI\/ML) will be the key pillars for the network of the future. It is one of the most exciting times to be in the communication industry with so much to innovate, so much to transform and so much to grow.<\/p>
<\/p>
Growing video consumption and new applications will continue to drive exponential data traffic growth on mobile networks; video content itself is getting heavier with 4K video. By 2021, mobile video data consumption is set to increase by 870% from 2016.Further AR\/VR devices will have a massive multiplier effect on the total data traffic volume and will put much higher \u2018throughput\u2019 demands on the network of the future. Massive IoT will disproportionately push the network limits on total number of \u2018connections\u2019 as the industry embarks to connect 20B devices and more.And new applications like driverless cars and self-flying drones will demand much lower network \u2018latency\u2019. Network of the future has to meet these unprecedented demands on: throughput, connections and latency.<\/p>
Some of you might recall the massive challenges CSP networks faced in the early days when smartphones were launched. Network engineers on some days were dealing with \u2018data tsunamis\u2019 as smartphones pushed the network\u2019s limits on data throughput. And on others days the same network engineers would be dealing with \u2018signalling storms\u2019, when un-optimized apps were creating havocs on network\u2019s signalling traffic. It is worth noting that networks experienced all these challenges with just one type of \u2018thing\u2019: smartphones. We can foresee what\u2019s to about come with IoT, when networks will have to deal with massively varying use-cases from massively varying devices. Massive IoT further exacerbates security threats, last year\u2019s Mirai DDoS botnet attack clearly exemplifies the threats network of the future will face. It will no longer be viable to profitably operate and secure the network of future with traditional tools and processes. Hyper-automation is the only way forward to meet the hyper-demands of IoT. Big Data, AI, Machine Learning, Analytics will all play a key role in driving hyper-automation of the network of future.<\/p>
Furthermore, for service providers to seize the emerging opportunities networks will have to deliver more-for-less. According to GSMA, from 2016 \u2013 2020 data traffic on mobile networks is projected to grow at 47% CAGR, while revenue is projected to grow just at 2.1% CAGR during the same period. The math couldn\u2019t be simpler; to keep their networks sustainably profitable, service providers have to find solutions that aggressively bend their cost curve. To achieve this, network of the future requires: new architecture. Traditional network appliance models no longer profitably scales to meet the future needs, linear-scale must give way to hyper-scale. Service providers need an architecture that enable networks to be elastic, programmable and agile. Software Defined Networking (SDN) and Network Functions Virtualization (NFV) are the forces that are shaping up the architecture for network of the future. Web-scalers have already proven building hyper-scale software-defined infrastructure leveraging COTS platform and merchant silicon. SDN\/NFV promise to do just that for operators; driving optimal resource utilization and enabling them to innovate their networks at the speed-of-cloud.<\/p>
We strongly believe the three key pillars of the 5th wave of networks are going to be:<\/p>
Spectrum is the lifeblood of a mobile network. To dramatically increase network throughput and significantly lower latency, network of the future requires new access. 5G NR is shaping up to address these requirements by significantly improving spectral efficiency and by leveraging higher frequency bands (>6GHz), including mmWave, where much higher channel bandwidths are available. Network of the future will leverage both licensed and unlicensed spectrum, as well as new shared spectrum options like: CBRS 3.5GHz in the U.S. Numerous 5G NR trials are already underway as service providers mature RF technology for 28GHz, 31GHz and even 60GHz mmWave. We are also witnessing dis-aggregation of the base station itself with vRAN (virtualized RAN). vRAN will dramatically lower the cost of the RAN equipment. RRH (Remote Radio Heads) are being decoupled from BBU (Baseband Unit), and BBU itself is being virtualized. And that brings me to NFV.<\/p>
<\/p>
SDN and NFV are laying the foundation of the network of the future by adapting the Cloud technologies for the carrier networks. We are already witnessing the growing importance open-source-software will play in the network of the future. The industry is humming with numerous open source projects including: OpenStack, OpenDayLight, ONOS, ONAP, OSM, CORD and more. The software architecture itself is evolving to microservices, as service providers demand solutions to be cloud native. Old stovepipe OSS stacks are starting to give way to new-age Management and Orchestration (MANO) solutions that will bring unprecedented agility, enabling service provider to quickly design, realize and offer new services. New architecture will enable service providers to achieve closed-loop-control, such that the network elastically scales up and down depending on varying traffic conditions. Finally, the new architecture will enable network slicing, enabling service providers to programmatically create network slices catering to varying demands of various IoT use cases.<\/p>
<\/p>
Pairing vast amounts of network data collected in data lakes with the vast amount of compute power now creates the potential to unleash the power of machine learning and AI algorithms to achieve unprecedented levels of hyper-automation. Future networks will be self-aware and self-optimizing. Leveraging predictive analytics these networks will achieve optimal asset utilization and near real-time auto-scaling. They will be able to predict faults and self-heal. Hyper-automation will pave the way for self-defending networks, that will be able to detect traffic anomalies and proactively prevent against security attacks like the recent Mirai massive DDoS Botnet attack.<\/p>
<\/p>
To conclude, the future of the communication industry looks very bright as the demand for data continues to grow. IoT presents unprecedented growth opportunity for service providers. To successfully seize these growth opportunities, the service providers must transform their networks to achieve unprecedented levels of agility, elasticity and automation. The 5th wave of networks will be go much beyond the Radio innovation. New Radio (NR), New Architecture (SDN\/NFV) and Hyper Automation (AI\/ML) will be the key pillars for the network of the future. It is one of the most exciting times to be in the communication industry with so much to innovate, so much to transform and so much to grow.<\/p>
<\/p>