What’s up with… Verizon & Open RAN, Tele2, US network security

• Verizon wades deeper into Open RAN waters
• Tele2 replaces its CTIO
• US concern about Chinese cyberattacks is mounting

In today’s industry news roundup: Verizon combines Samsung and Commscope tech at some toe-in-the-water Open RAN deployments; Yogesh Malik is looking for a new job; fears over mounting cyberattacks, especially against telecom networks, are reaching fever pitch in the US; and much more!

US telco giant Verizon has taken its latest step into the disaggregated, multivendor mobile network technology waters with the news that it has deployed Open RAN-based distributed antenna systems (DAS) systems at The University of Texas Moody Center and the Austin Convention Center. The deployments, which Verizon says mark “a major step towards larger-scale, multivendor deployments using Open RAN,” comprise Samsung Networks virtual distributed units (vDUs) with open interfaces to DAS radio access technology supplied by Commscope. According to the operator, the move shows how the use of Open RAN gear “can bring together best-in-class products from the ecosystem to create a highly efficient solution and provide Verizon’s high-performing 5G Ultra Wideband service throughout the venues. Not only does this commercial deployment mark a major step towards larger-scale, multi-vendor deployments using Open RAN, but it materially impacts the cost of power, space and cooling by eliminating unnecessary radio frequency equipment, and improves reliability resulting in a gold-standard customer experience for fans and visitors to these venues.” It added: “The transition to Open RAN has the potential to bring many benefits in terms of deployment flexibility, faster innovation in an open environment, and greater service options by increasing the opportunity for new entrants to provide competitive and advanced solutions. More competition, more innovation, and increased supplier diversity will all be net benefits to operators and customers.” That Verizon is testing the waters with Open RAN deployments, even at this small scale, shouldn’t come as a surprise to anyone following the operator’s network strategy as it has noted from the beginning of its engagement with Samsung Networks in 2020 that it would be broadly deploying Open RAN-ready virtual RAN (vRAN) systems from the vendor and set itself a target of having 20,000 sites based on Samsung’s technology by 2025, In fact, early this year it noted it had reached 15,000 such sites and deployed more than 130,000 Open RAN capable radios in its network. Adam Koeppe, senior VP of technology planning at Verizon, noted: “The massive evolution of our network over the past few years, including our move to a cloud-based architecture, widespread virtualisation and our aggressive adoption of O-RAN standards and capabilities has enabled us to show O-RAN interoperability success in a commercial environment.” Further such action is expected at Verizon following the recent appointment of major Open RAN advocate Santiago ‘Yago’ Tenorio as its CTO and senior VP of strategy and technology enablement. Verizon, of course, isn’t the only major US operator talking up its Open RAN activities at the moment…. See Rob Soni on AT&T’s Open RAN strategy.

Yogesh Malik has suddenly stepped down from his role as CTIO at European network operator Tele2, which has operations in Sweden, Estonia, Latvia and Lithuania, and has been replaced with immediate effect, with the operator’s current head of digital capabilities and technology for Sweden, Ove Wik, taking over as acting CTIO. Tele2 claims that after more than three years at the company, Malik had successfully completed the operator’s 5G rollout and IT migration, so his job was done! But that is clearly a smokescreen and it’s worth remembering there has been a lot of change this year at Tele2, with Xavier Niel’s Iliad Group emerging as a major investor at the beginning of this year, the subsequent appointment of Iliad CEO Thomas Reynaud as chairman and, just recently, the appointment of a new CEO in the form of Jean Marc Harion, previously the head of Iliad’s operations in Poland. Such top-level changes always lead to reshuffles further down the chain and it looks to the team here at TelecomTV that Malik’s departure is part of what might be an ongoing senior management-level revamp at Tele2.   

The US is getting more and more concerned (and angered) by the ongoing spate of hacking attacks on the nation’s strategic communications infrastructures to the extent that the director of the National Security Agency (NSA), Timothy Haugh, who is a four-star general in the US Air Force, is pressing the businesses and organisations across the huge American private sector to quickly submit to the NSA full details of any and all attempts to breach their comms and data systems by ‘bad state actors’ in general, and by Chinese hackers in particular. General Haugh also has responsibility for US Cyber Command, a unit of the Department of Defense dealing with military, intelligence and IT capabilities. Its mission is to direct, synchronise, and coordinate cyberspace planning and operations to defend US national interests in collaboration with domestic and international partners. There can be absolutely no doubt that states such as Russia, Iran and North Korea continually attempt to hack into US systems, but China is the most determined nation to target, infiltrate, spy on and disrupt the comms of the US and its allies. Just take a look at this log of significant cyber incidents produced by the Center for Strategic & International Studies: There are far too many to cite here but the number, depth and breadth of attacks is astonishing. Speaking at the National Security Innovation Forum in Washington DC, General Haugh told Bloomberg that the Chinese state is attacking US comms assets “en masse” and as a national effort, adding that “public disclosure would help find and oust the hackers”. To help, he intends to compile a public “hunt guide” to enable companies and their cybersecurity specialists to find where and how hackers are working and then “eradicate them” from telecom and data networks. General Haugh added, “The ultimate goal would be to be able to lay bare exactly what happened in ways that allow us to better posture as a nation and for our allies to be better postured.” That phraseology might sound rather like an extract from an advertorial for a nationwide lumbar support programme, but the problem is urgent. Only this week Richard Blumenthal, a Democratic senator for the state of Connecticut, urged rapid action to counter the “sprawling and catastrophic” infiltration being inflicted on US comms systems, including those of big and strategically vital telcos such as AT&T, Verizon and T-Mobile US. General Haugh said that the availability of “detailed public disclosures” could help enterprises and organisations that have not yet experienced incursions (or have not yet discovered that they have been compromised) to “begin to put countermeasures in place.” The system would also “help other nations uncover and root it out too.” He concluded, “It’s going to take collective work” – and then some. The Chinese authorities in Beijing and at the embassy in Washington DC flatly deny that the PRC is involved in any programmes to infiltrate and disrupt US comms networks. 

General Haugh’s comments came as US Senator Mark Warner, who is chair of the Select Committee on Intelligence, told the Washington Post that the recent breaches of US telcos by China-affiliated hackers constituted the “worst telecom hack in our nation’s history – by far”, reported Reuters. 

Belgium’s national operator, Proximus, is raising €105m from the sale of its towers portfolio in Luxembourg to infrastructure asset management firm InfraRed Capital Partners. Proximus Luxembourg, which operates under the Tango and Proximus NXT brands, will remain an anchor tenant on the tower sites, of which there are 267. “This agreement represents another milestone in our bold2025 strategy to unlock value through asset divestments,” noted Proximus CEO Guillaume Boutin. “By partnering with InfraRed Capital Partners, we ensure the long-term stability and operational continuity of our mobile infrastructure in Luxembourg, while freeing up resources to support transformative growth projects like our fibre roll-out strategy.”

Aviz Networks, which has developed a software stack that supports the SONiC open-source network operating system (but which describes itself as an AI-driven networking solutions company), has raised $17m in its Series A funding round. The round was led by Alter Venture Partners and included existing backers Cisco Investments, Moment Ventures, Wistron and Accton as well as two new investors in the form of tech manufacturing giant Celestica and Qualcomm Ventures. The fundraising has been ongoing for a while – with the round reaching $10m about a year ago. The company plans to increase its workforce both in the US and globally, and expand into new sectors with its networking software that aims to meet “the evolving demands of datacentres and edge networks as they scale and integrate AI”. Aviz’s CEO, Vishal Shukla, noted: “Traditional networking solutions, even the best, struggle to adapt as infrastructure increasingly depends on AI and data-driven applications. A vendor-agnostic approach is crucial to advance networks to the next level. This strategy leverages top-tier networking hardware, a robust network operating system designed for AI, and end-to-end networking tooling that integrates AI directly into everyday networking. Aviz has perfected this approach, placing AI at the heart of how networks are built and managed. This funding round will boost our growth in current markets and allow us to explore new verticals, strengthening our commitment to innovation and excellence in an AI-driven era. We are excited to welcome new investors and deeply appreciate the ongoing support from our existing investors, partners & customers.” 

The ever-increasing popularity of online multiplayer gaming is having a massive knock-on effect in terms of the need for speed and bandwidth. This week, the demand hit such a level that Europe’s biggest internet exchange and data hub, DE-CIX (Deutsche Commercial Internet Exchange) in Frankfurt, Germany, broke the record for data throughput when at 23h11 on Wednesday evening it was shifting 18.11 Tbt/s, the exchange reported. That, apparently, is the equivalent of streaming 10 million TikTok videos simultaneously. Be still my beating heart. The sudden upsurge in demand was hardly coincidental. Several big companies had either published new games or updates to established releases and players responded with due enthusiasm, knocking back fizzy drinks and munching on packets of crispy snacks late into the night, with bedtime deferred as caffeine and E-numbers coursed through their veins. Over the past year, traffic at DE-CIX has grown by 9.6% and since 2020 has risen by 118%. The company operates carrier- and datacentre-neutral internet exchanges, with operations in Europe, North America, Africa, the Middle East, India and South-east Asia. Ivo Ivanov, CEO of DE-CIX commented, “Digital and high-bandwidth, data-based services are driving the growth in traffic at Frankfurt”. [However,] “internet exchanges are not only at the heart of all online offers for end users when it comes to streaming and gaming but also for industry, which can use them to open up new intelligent, data-driven business models.”

You might think that the world’s accelerating descent into demented dystopia is already bad enough, but think again. Apparently AI can now create a ‘replica’ of your personality. As reported by the inestimable MIT Technology Review, all it takes to clone your personality “with an 85% degree of accuracy” is for you to sit down for a spoken interview with an AI model chatbot, during which you will be prompted to tell it all about your “values and preferences”, which will then be “captured” by the behavioural simulation machine. The interaction takes two hours (that you’ll never get back) and once it is over and the material is loaded onto “the system” you can be split into two parts: One your physical self, and the other a virtual replica that, in turn, can itself also be split into an infinite number of mini-me digital replicas. The research is underway at Stanford University in California and the UK/US DeepMind AI laboratory that, since 2014, has been a subsidiary of Google. In that time, DeepMind has developed neural Turing machines resulting in a computer that can, in many ways, resemble the short-term memory of the human brain. Back in 1949, British computing pioneer Alan Turing devised his eponymous ‘Test’ (he called it “The Imitation Game”) to determine a machine’s ability to exhibit intelligent behaviour equivalent to, or indistinguishable from, that of a human. The Stanford/DeepMind research, which is entitled ‘Generative Agent Simulations of 1,000 People’, is authored by a team of nine scientists led by Joon Sung Park. It is available on arXiv, the open archive for academic articles that is maintained and operated by Cornell University of the US. The material posted on arXiv has not been peer-reviewed. Not scary enough so far? Read on. The Stanford team recruited 1,000 people. They varied by age, gender, ethnicity, race, levels of education, and political ideology. They were paid up to $100 each to bare the innermost recesses of their personalities. Following the interview, “agent replicas” of correspondents were created. To determine how well the replicas reflected their human counterparts, each of the human guinea pigs took a series of personality tests and social surveys, and also participated in logic games. Each volunteer took them twice, at fortnightly intervals. Thereafter, the replicas did the same exercises as the humans and it was found that they mimicked the unique personality traits of those subjects 85% of the time. By the way, the word ‘replica’ has connotations in that it is uncomfortably close to the android ‘replicants’ of the seminal science-fiction film ‘Blade Runner’, so now they are called ‘simulation agents’ instead. So much more human, don’t you think? It is claimed that one of the reasons why they have been created is to make it easier for those working in various fields to research and produce studies “that would be expensive, impractical, or unethical to do with real human subjects. If you can create AI models that behave like real people, the thinking goes, you can use them to test everything from how well interventions on social media combat misinformation to what behaviours cause traffic jams.” Hmm. Joon Sung Park, naturally, is very enthusiastic about the prospect of replicas. He says, “If you can have a bunch of small ‘yous’ running around and actually making the decisions that you would have made – that, I think, is ultimately the future.” Well, at best, 85% of the time anyway. Not everyone will be quite so sanguine. Thanks to AI-aided image generation technology, cyberspace is already riddled with deep fakes and deep-fake content. It surely wouldn’t take long for criminals and megalomaniacs to build the tools to present perverted personalities of real people online and have them saying things and recommending things that the original person would (hopefully) never articulate. In these troubled times, we might do well to remember the story of the opening of Pandora’s box.

– The staff, TelecomTV