Johannesburg, 24 Nov 2020
By 2025, there will be 100 billion connections and 6.2 billion people will have access to the Internet. This is according to the Huawei Global Industry Vision 2025, which also predicts that 85% of enterprise applications will be cloud-based.
We are already witnessing and will continue to see, the exponential growth of enterprise cloudification and digital transformation. With this growth comes an increasing number of key services that IP networks need to carry – giving rise to many new network challenges.
Some of these challenges have been outlined by Gartner Research’s VP, Andrew Lerner, whose team examined key technologies that will promote enterprise infrastructure and operations (I&O) in the future.
Given the enterprise’s almost total reliance on the network in the future, intelligent networking will necessarily become the de facto solution that supports industries.
AIOps, big data, ML
To this end, Lerner’s team identifies artificial intelligence for IT operations (AIOps) as a key technology. “To be more specific,” says Peter Ye, Data Communication Product Line, Huawei Technologies, “it uses AI to provide full visibility into the status and performance of IT systems on which enterprises depend. The AIOps platform combines big data and machine learning to analyse a large amount of data from multiple data sources and provide multiple analytical and presentation technologies.”
Gartner predicts that by 2022 the 2% (2018) of large enterprises deploying AIOps platforms will have mushroomed to at least 25%.
“We believe,” says Ye, “there is no time to waste when it comes to building AIOps platforms for enterprise network operations and maintenance (O&M).”
In an AIOps versus automatic O&M scenario, AIOps:
- Solves more problems;
- Provides a more intelligent basis for ICT O&M decision-making; and
- Predicts trends, allowing for stable operation and growth.
Performance and power efficiency
Says Ye, performance and power efficiency will become increasingly critical, with compute accelerators standing front and centre as AI loads computational resources. Gartner estimates that by 2022, computational resources used in AI will be four times more than in 2018.
From graphics processing unit accelerators to DNN application-specific integrated circuits, and field-programmable gate array accelerators, compute accelerators can greatly improve the computing performance,” he says.
“However, if there is packet loss on a traditional computing network, it will not be able to reach its full potential in terms of computing power. As such, it has become the bottleneck for improving computing power in the AI era. To fully unleash the computing power of data centres, enterprises urgently need to introduce the zero-packet-loss network.”
Distributed computing
EC-IOT, or edge computing Internet of things, places information processing close to the things or people that produce or consume that information. As such, says Ye, latency and unnecessary traffic are reduced, as processing is kept local.
“Edge computing solves many pressing issues, such as excessive latency, insufficient bandwidth, and high costs, helping to cope with the massive increase in edge-located data as positioning applications become increasingly popular,” says Ye.
Ye says Huawei’s prediction is that edge computing poses new requirements on network intelligence, latency, bandwidth and access. As a result, it will pressure the transformation of network technologies from lossy to lossless, from “best-effort” to “deterministic”, from dumb traffic pipes to intelligent computing networks, and from limited access to access anytime, anywhere.
Reducing opex while optimising performance
Ye says that as pressure mounts for networks to increasingly improve capacity and efficiencies, intent-based networking systems (IBNS) will become a serious consideration for large enterprises. “IBNSes improve network agility and availability and support unified intent and policy across heterogeneous infrastructures. When the technology matures, a full IBNS implementation will reduce the time to deliver network infrastructure services to business leaders by 50%-90%,” he says.
“It will also reduce the number and duration of outages by at least 50%. In addition, IBNSes reduce operating expenditure, optimise performance, cut dedicated tooling costs, enhance documentation and improve compliance.”
This technology also provides:
- Translation and validation, by taking a higher-level business policy as input from end-users and converting it to the required network configuration;
- Automation, by configuring appropriate network changes across an existing network infrastructure;
- State awareness, by ingesting real-time network status for systems that it controls; and
- Assurance and dynamic optimisation – by continuously validating that business intent is being met, it can take corrective action when it isn’t.
While there are fewer than 15 large enterprises utilising IBNSes today, Gartner’s prediction is that by 2022, more than 1 500 will use it.
Networks for enterprises of the future
Users need more secure, intelligent and agile networks. Various innovative technologies are emerging and digital transformation requirements of various industries impose increasingly strict requirements on networks.
“At Huawei, we’re helping customers to build ubiquitous ultimate connection experiences and accelerate enterprise digital transformation with our IDN solution. Driven by technology and customer requirements, Huawei continuously invests in and innovates to build better ICT solutions, maximise customers’ business value and lead the intelligent IP era,” Ye concludes.
Share