Networks in an AI world

07 February 2024

The future of networks is on an unknown trajectory amidst paradigm-changing new technologies like AI shaking up the market. We speak to the experts to find their views for 2024 and beyond…

Enter chief AI officer…
There’s no denying that 2023 was the year of artificial intelligence (AI). Hitting the headlines of all major IT/networking publications throughout the year, the impact of AI on the networking and IT space is only beginning to make itself felt.

Still in its infancy, Apricorn’s managing director, EMEA, Jon Fielding, believes that 2024 will see more use of AI assisted data management, aiding businesses to automatically classify data according to whether it is relevant, valuable, or risky, to decipher which they should retain or remove, greatly assisting data retention strategies.

“In doing so, the business can conserve costs by saving a huge expense on unwarranted data storage space. Using AI to automate the process of classifying data also improves the efficiency of data compression with AI algorithms, again limiting the required storage space and the time needed to retrieve stored data,” explains Fielding.

“Those businesses that look to embrace the use of AI for data management in the coming year will reap the cost and performance benefits to data storage systems in the long term, as well as bettering their data security defences.”

However, “despite all the hype around AI and generative AI, the technology is far away from being able to automate and optimise every aspect of our lives anytime soon,” says Mark Pierpoint, VP of strategic innovation and partnerships, Keysight. “AI is making progress; however, automating a chatbot or creating a digital assistant are constrained problems that are much easier to automate. When it comes to helping manage real-world processes such as optimising call quality on a 5G network or managing energy consumption, these are incredibly complex operations, with a wide array of variables requiring vast unbiased data sets before AI can be effective.”

Jim Liddle, chief innovation officer at Nasuni, asserts that the rush for an AI advantage is surfacing deeper data infrastructure issues that have been mounting for years: “AI doesn’t work in a vacuum and it’s just one part of the broader data intelligence umbrella,” he says.

“Given the rapidly changing landscape, a pressing priority for businesses to consider in 2024, and one highlighted by Frank de Jong, programme director edge computing, CTIO office, Orange Business, is AI in the C-Suite and at board level: the role of chief AI officer has just emerged.”

Many organisations have implemented data analytics, machine learning (ML) and AI into their sales, customer support, and similar low-hanging initiatives, but struggle to integrate the technology in more sophisticated, high-value applications.

“Visibility, for example, is a crucial and often-overlooked first step towards data intelligence,” adds Liddle. “A shocking number of companies store massive volumes of data simply because they don’t know what’s in it or whether they need it. Is the data accurate and up to date? Is it properly classified and ‘searchable’? Is it compliant? Does it contain personal identifiable information (PII), protected health information (PHI), or other sensitive information? Is it available on-demand or archived? In the coming year, companies will be forced to come to terms with the data quality, governance, access, and storage requirements of AI before they can move forward with digital transformation or improvement programmes to give them the desired competitive edge.”

Although AI is providing a competitive advantage for those that can leverage it effectively, Thomas King, CTO at DE-CIX, explains that many companies lack the internal resources to enable the development and operation of their own AI models.

“Here, AI from the cloud and AI as a Service (AIaaS) providers offer convenient alternatives. For these, optimised connectivity to clouds and specifically to the AIaaS providers is an essential component for ensuring low latency and high bandwidth, and thus better performance of chatbots, analytics, and other tools,” says King.

Given the rapidly changing landscape, a pressing priority for businesses to consider in 2024, and one highlighted by Frank de Jong, programme director edge computing, CTIO office, Orange Business, is AI in the C-Suite and at board level: the role of chief AI officer has just emerged.

“AI technology developed rapidly through 2023 improving efficiencies in many departments within businesses. I predict that the speed of development will be maintained at a similar trajectory, and even increase in tempo, throughout 2024, which means that the time for businesses to internally consider the ethical development of AI technology is now,” says de Jong. “Orange Business urges enterprises to put in place someone who can champion and lead this development from within… Things are going to move fast.”

AI holds great promise for organisations of all sizes this year, but change won’t be instant, nor necessarily even that fast. Industry experts believe that we’ll be looking at several years of pilots, tests, and upgrades while those in the value chain get to grips with all that AI has to offer.

“While the intelligence will undoubtedly help us in 2024, realistically, AI will not be ready to direct physical-world activities until the end of the decade,” opines Pierpoint.

The changing face of storage
Data storage faces several significant challenges/opportunities coming into 2024 and beyond; capacity, security and, of course, AI.

“Many organisations that are reliant on data centres are reporting that their most pressing issue right now is one of capacity. A growing number of data centres are full, and don’t have the space or power available to deploy new platforms,” reports Fred Lherault, field CTO EMEA/emerging markets, Pure Storage.

This will result in widespread efforts to achieve efficiency gains to reclaim space and power to accommodate the use of new technologies. “We’ll see operators looking to switch to new, more power efficient technology, with smaller space and cooling requirements; in essence extending the life of the data centre - an essential factor when considering the need for new technologies in the wake of the rise of AI,” adds Lherault.

Moreover, as data generation continues to explode in line with digital transformation and AI adoption, so does its lifespan.

“At the end of the day, your data will outlive you,” says Tim Sherbak, enterprise products and solutions marketing, Quantum. “This means your organisation must have a plan for its long-term management and ongoing platform infrastructure today. How can it be easily curated, accessed, and analysed in the future when the current experts are no longer present, plus, the need to ensure that data is always durably protected. This includes implementing an evolutionary storage architecture that can seamlessly adopt and decommission across multiple generations technologies and platforms for decades to come at a cost that doesn’t break the bank. It’s no longer just an option to consider the ‘far future’ of your data – it’s a requirement for organisations focused on a fully digital-enabled future.”

Data storage security, too, remains an ongoing but increasing concern: “I expect we’ll see a growing trend in organisations moving data away from the public cloud globally in the next 12 months. It’s also something that will drive budget decisions, as IT leaders determine whether to apportion budget from the cloud to use instead for on-premises storage for certain data types,” says Guillaume Crapart, senior director of channel sales, Quantum. “The semi-private cloud is another storage option I believe we’ll see rise in popularity in 2024, as organisations seek to have their cloud-based data closer to home and subjected to stronger security measures than the public cloud allows for. In combination with the cloud-also trend, organisations will gain more privacy and control over their data - a must in today’s digital world.”

Interestingly, the rush to develop and implement generative AI solutions is leading to a new wave of adoption of containers and Kubernetes from data scientists and developers.

“There’s a growing realisation that containers are fundamental to AI at almost all stages, in that every AI tool is packaged in containers,” says Patrick Smith, field CTO EMEA, Pure Storage. “They are also vital to the success of the hard part of AI – training large language models. Before a model can be trained, many processes must be performed, including data curation, cleaning, and amplification. The tools that perform these tasks all reside in containers.”

According to Smith, there have been two waves in the rise of containers and their evolution: “the first was stateless, which involved workloads going into containers, the second was stateful, with databases and critical applications moving to containers, requiring data storage. We’re now entering a third wave, with containers driving and enabling the AI data pipeline and toolchain.”

Meanwhile, with AI already transforming data centres, huge investments are being triggered by cloud service providers to build out AI fabrics to keep pace with demand.

“In 2024, look for intense focus and investment in GPUs and other back-end infrastructure for large learning clusters,” says Stephen Douglas, head of market strategy, Spirent. “Cloud service providers will also be looking to beef up front-end inferencing and investing in faster interconnect technologies to enable it. We expect Ethernet to dominate these front-end networks, and for 800G adoption to grow rapidly among Tier-1 cloud service providers next year as 51.2Tbps switches reach the market.”

Indeed, global spending by businesses on cloud is forecast to top $1 trillion for the first time in 2024, driven by factors like adopting new tools such as AI.

“The cloud computing industry is set to change significantly. A trend expected to see gaining momentum is the adoption of software as a service (SaaS) public cloud offerings as part of next-generation ERP systems,” shares Don Valentine, commercial director, Absoft. “This shift builds upon the recognition that merely moving existing ERP systems to the cloud does not automatically trigger innovation. Instead, it is the combination of cloud infrastructure with cutting-edge ERP products that propels businesses toward digital transformation.”

“Interestingly, the rush to develop and implement generative AI solutions is leading to a new wave of adoption of containers and Kubernetes from data scientists and developers.”

The surge in AI adoption has so far had a limited impact on the cloud computing market, “however, it is expected to have a more noticeable influence once the offerings from major hyperscale cloud providers become more widely available and commercialised,” says Lee Thatcher, head of cloud, CloudCoCo. “As these providers continue to invest in AI capabilities, businesses will increasingly leverage cloud platforms for AI-related workloads, driving the growth of cloud-based AI applications and services. Therefore, it’ll continue to be about AI for the foreseeable future.”

Mergers, acquisitions, and subscriptions
According to all reports, companies are waking up from the post-covid hibernation, with much more M&A activity expected in the coming 12 months.

“The days of only working with the big globally recognised vendors, who likely have been the mainstay in the security stack long before the current security team was formed, is gone,” says Bolton. “As a result, these older and more established companies are looking to acquire these start-up and scale-up companies to absorb the ingenuity held within them and repackage it as a differentiator for themselves before their value gets too high.”

Despite this, continued concerns about the economy, difficult business conditions and high energy costs will have a significant influence in 2024, accelerating the trend towards OpEx spending over CapEx.

“Money is tight and continued belt-tightening in the coming year will make technology subscription services very attractive as customers only want to pay for what they use, avoiding the need for large CapEx outlays,” agrees Lherault. “It makes sense to opt for a subscription to a service and avoid CapEx expenditure when there’s a reasonable possibility that a CapEx asset will not be used to full capacity.”

The ‘as a service’ model will, however, only gain traction if backed by relevant SLAs, since buyers are becoming increasingly discerning about what to invest in.

“This trend is also being observed in data centres, with operators beginning to favour on-demand models that enable just-in-time provisioning of assets,” says Lherault. “This allows them to better control energy costs through lower power consumption, which also helps them realise their sustainability goals.”

All in all, 2024 looks like a year of great change for the IT community. AI continues to bring huge advances – but also huge threats – to all aspects of the network, everywhere, changing and challenging the operating environment like never before.