News

Cloudflare’s 1.1.1.1 DNS resolver service fell victim to a simultaneous BGP hijack and route leak event, causing massive internet outages and degradation worldwide. Pakistan caused the most famous BGP outage. The government tried to block access to YouTube within the country. Their misconfiguration caused a worldwide YouTube outage.

Most organizations are targets of attacks 7.5 times a year. And while most are resolved quickly, these are examples of public infrastructure failures that are beyond your control.

What other technology do you rely on every day that was invented in the 1980s? Not your smartphone. Not your car. Not your TV. And definitely not your work tools. Yet, every time you send an email, connect to a website, or deploy a cloud service, you’re relying on core internet protocols that predate the web itself.

The Border Gateway Protocol (BGP) was designed in 1989, an era when the “internet” was barely a concept and security was an afterthought. Back then:

- Home users connected via dial-up modems.

- Businesses considered themselves cutting-edge if they had a T1 line.

- Network reliability was a hope, not an expectation.

BGP’s original purpose was simple: keep the nascent internet stitched together. It provided large institutions with a means to announce which IP address blocks they controlled and to learn about others. The protocol allowed routers across autonomous systems (ASes) to share route announcements and dynamically discover paths to distant networks.

BGP was designed for resilience, not determinism. For openness, not security.

Today, we demand speed, uptime, and security that BGP was never built to deliver. Multi-gigabit fiber reaches homes. Enterprises span multiple clouds across continents. Workloads like real-time video, financial transactions, and machine learning require low-latency, high-throughput data paths.

However, BGP still routes traffic based on trust and reachability, rather than performance or identity. It can’t enforce policies. It can’t prevent hijacks. And it certainly can’t guarantee who’s on the other end.

Despite multiple security incidents and efforts, such as RPKI and BGPsec, the internet still routes traffic based on a chain of trust that can be exploited by anyone with a few malicious route announcements. Most fixes require coordination that doesn’t exist and IT infrastructure upgrades that move at glacial speed.

The result? The modern internet rides on a protocol that thinks it’s still 1992.

Another artifact of that era is the Domain Name System (DNS). Created to make numeric IP addresses human-readable, DNS transformed how people accessed websites. Instead of memorizing strings of numbers, you could simply type in a name.

The problem? DNS is public by design.

Every query, every resolution, and every domain is visible and discoverable. Attackers can enumerate subdomains, discover shadow IT resources, and probe for vulnerabilities – all by posing as legitimate users.

We’ve seen this pattern before. Consider phone numbers. In the 1990s, receiving a call or piece of mail felt like an event. Now? Most calls are spam, and most email is junk. People don’t pick up unless they recognize the number. Our relationship with public identifiers has undergone a fundamental shift.

The same evolution is happening with network services. Public IP addresses and DNS names are easily scraped, scanned, and attacked. In an age of automation and AI-assisted hacking, exposing your infrastructure by default amounts to sending an invitation.

Yet we continue treating server addresses like phone numbers in a white pages directory – a model that no longer works for the threats we face.

Both BGP and DNS reflect assumptions that simply don’t hold up anymore:

- Assumption: Networks are trusted.

-- Reality: Most attacks now originate from within or via compromised peers.

- Assumption: Routes are stable.

-- Reality: Internet routes change unpredictably due to performance tuning, outages, and misconfigurations.

- Assumption: Identities don’t matter.

-- Reality: Zero-trust architecture has become the standard for secure design.

- Assumption: Services are few and fixed.

-- Reality: Modern architectures dynamically spin up and down thousands of services.

The more we scale and automate, the more these assumptions crumble.

The internet’s early architecture was undeniably brilliant for its time. But that time has passed.

Today’s needs are different. We need:

- Deterministic data paths that can be trusted end-to-end.

- Secure naming systems that are private by default.

- Policy-aware routing that aligns with business, performance, and compliance requirements.

- A model where services announce themselves securely to authorized peers, not to the entire internet.

These aren’t enhancements; they’re necessities.

The irony is striking: everything else in tech has evolved dramatically. Compute became elastic. Storage turned redundant and distributed. Deployment went fully automated. But networking? It’s still largely manual, primarily public, and built mainly on 40-year-old concepts.

This should be our wake-up call. We can’t keep patching internet security with duct tape and hoping for the best. It’s time to challenge the status quo and ask a hard question: are the foundational protocols we depend on every day actually fit for purpose anymore?

Security and privacy can’t remain afterthoughts we layer onto a crumbling foundation. They need to be built from the ground up. That means completely reimagining how the internet connects, routes, and identifies everything.

Think about it: what other critical system in your life still runs on ideas from the 1980s?

LINK!

This article was produced as part of TechRadarPro's Expert Insights channel where we feature the best and brightest minds in the technology industry today. The views expressed here are those of the author and are not necessarily those of TechRadarPro or Future plc. If you are interested in contributing find out more here: https://www.techradar.com/news/submit-your-story-to-techradar-pro

Congratulations, fellow Stranger Things fan! You’re on the home stretch of the three year wait between Stranger Things season 4 and Stranger Things season 5. The final episodes will be split into three releases – volume 1 on November 26, volume 2 on December 25 and the season 5 finale on December 31 – so there’s still a small wait to go, but rumour has it we’re getting the first full season 5 trailer at some point this week.

Today (July 15) marks the first time we ever saw one of the best Netflix shows of all time on screen, with the series debuting nine years ago in 2016. If you can’t remember what happened when we last visited Hawkins (and that’s understandable), our group of best friends attempted to defeat Vecna, causing Max's apparent death as well as the opening of a massive rift between the town and the Upside Down. No big deal, I’m sure.

But as Netflix finally tries to get its fanbase excited about the drawn-out end, a painful question has to be asked. Why should anybody care about Stranger Things season 5 when we’ve been left in the lurch for so long? I’m wondering if it would have been less of a hassle to have been eaten by Vecna when we first met him, and that’s a problem.

Let’s put it into context. Since Stranger Things season 4 aired, we’ve had four seasons of The Bear, five seasons of Slow Horses (if you count new episodes we’re going to get in September) and two seasons of Severance… and look how long that took to return. In the Stranger Things world alone, we’ve had the arrival of non-canon West End play The First Shadow, plus the announcement of two spinoffs: animated series Stranger Things: Tales from '85 and live-action show The Boroughs. Everyone and their nan has seemingly complained about not getting season 5 in the meantime, and they’ve got good reason to be annoyed.

Back in the good old days of the mid-2000s, we were regularly whipping through 22-episode seasons of TV like there was no tomorrow. Desperate Housewives and Lost were great examples of this, each requiring a high level of input and resource in their own way. Fast forward two decades, and the consolation prize of feeling lucky to get eight new episodes in three years doesn’t feel like something worth investing in.

Sure, these upcoming episodes are basically feature length movies and the technical craft needed to achieve them is immense, but this is Hollywood, for goodness sake! Every resource we allegedly have at our disposal is supposed to be at the top of its game, able to give us everything we have and haven’t yet dreamed up. From a marketing perspective, Netflix might have thought dragging out the jewel in the crown of its streaming back catalog would make fans hungrier for the end product, but there’s only so far you can stretch the theory in practice.

Of course, I’ll be streaming Stranger Things season 5 like the rest of us, but it will be a reluctant watch. The endless wait over the last few years has definitely made me think twice about investing in shows on one of the best streaming services around, and that’s before we even touch on the frequent cancellations (another story for another day).

• Stranger Things season 5's 12-month shoot yielded 650-plus hours of footage for its eight 'blockbuster movie' episodes, the popular Netflix show's creators say
• Stranger Things: The First Shadow's big lore reveal needs addressing in hit Netflix show's final season
• Everything new on Netflix in July 2025

• AI job losses are inevitable, but new innovations will curb the effects, Nvidia CEO Jensen Huang says
• Huang has been warned by senators before an upcoming trip to China
• US tech should 'set the global standard' Huang argues

It’s long been prophesied AI will lead to mass unemployment, with several CEOs and tech leaders warning AI will wipe out millions of jobs, and firms such as Microsoft laying off thousands of workers whilst bringing in new AI productivity tools.

Now, Jensen Huang, CEO of chip manufacturer and AI firm Nvidia, offered his (slightly stale) perspective. In an interview with CNN, Huang essentially passes job protection responsibilities over to business leaders, claiming; “If the world runs out of ideas, then productivity gains translates to job loss.”

“Everybody’s jobs will be affected. Some jobs will be lost. Many jobs will be created and what I hope is that the productivity gains that we see in all the industries will lift society,” Huang said.

Huang’s authority on AI is significant too, thanks to Nvidia's power in the market. The company's GPUs remain one of the most influential tech products in the world, and are largely powering AI development across the world - including in China, which is spooking some US politicians.

Huang recently received a warning written by Republican Senator Jim Banks and Democratic Senator Elizabeth Warren, Reuters reports, which advised against meeting with Chinese companies, arguing this could, “legitimize companies that cooperate closely with the Chinese military or involve discussing exploitable gaps in U.S. export controls”.

A Nvidia spokesperson saidUS technology will ‘set the global standard’ and that ‘America wins’ - with China being one of the largest software markets in the world, adding that AI software "should run best on the U.S. technology stack, encouraging nations worldwide to choose America”.

That being said, Huang has recently argued Chinese military branches will avoid using US technology because of the associated risk; “it could be, of course, limited at any time” he argued, “they simply can’t rely on it”.

He added how Chinese military services, which are already developing powerful tools, “don’t need Nvidia’s chips, certainly, or American tech stacks in order to build their military.”

This comes in response to growing concerns that Chinese companies and military agencies will use US tech to enhance capabilities.

Increasingly harsh restrictions have limited China’s access to top AI technologies, aimed at curbing China’s tech and AI advancement - but concerns remain about the threat to US national security should China use US companies to develop its capabilities.

• Take a look at our picks for the best AI tools around
• Check out our choice for best antivirus software
• China has spent billions of dollars building far too many data centers for AI and compute

• Binarly spotted multiple flaws in UEFI firmware built by AMI
• AMI released fixes months ago, so users should update now
• Many Gigabyte motherboards reached EOF and thus won't be patched

UEFI firmware on dozens of Gigabyte motherboards is vulnerable to a handful of flaws which theoretically allow threat actors to deploy bootkits on compromised devices, establish stubborn persistence and execute additional malicious code remotely, experts have warned.

Security researchers Binarly recently discovered four vulnerabilities in UEFI firmware developed by American Megatrends Inc. (AMI). All four have a high severity score (8.2/10), and can lead to privilege escalation, malware installation, and other potentially destructive outcomes. They are tracked as CVE-2025-7026, CVE-2025-7027, CVE-2025-7028, and CVE-2025-7028.

Binarly reported its findings to Carnegie Mellon CERT/CC in mid-April 2025, resulting in AMI acknowledging the findings and releasing a patch in mid-June. The patch was pushed to OEMs privately, but apparently Gigabyte did not implement it at the time.

There are apparently more than 240 motherboard models that are impacted by these flaws.

Many won’t be patched at all because they have reached end of life, and as such, are no longer supported by Gigabyte. Instead, users worried about the vulnerabilities should upgrade their hardware to newer, supported versions.

Products from other OEMs are also said to be affected by these flaws, but until a patch is applied, their names will not be publicized.

UEFI firmware is low-level code that runs beneath the operating system, and whose job is to initialize the hardware (CPU, memory, storage), and then hand off control to the OS. When this code has flaws, threat actors can exploit them to install so-called “bootkits”, stealthy malware that loads at boot time, before the OS.

Because they run in privileged environments, bootkits can evade antivirus tools, and even survive OS reinstalls and disk replacements. This makes them highly persistent and dangerous, especially in high-security environments. The good news is that exploiting these vulnerabilities often requires admin access, which is not that easily obtainable.

Via BleepingComputer

• The first UEFI bootkit malware for Linux has been detected, so users beware
• Take a look at our guide to the best website builders around
• We've rounded up the best password managers

• Nintendo has introduced new guidelines for the Switch 2 eShop in Asia
• The guidelines target several topics, including game bundles, and how they can be sold, sensitive content restrictions, and more
• These new guidelines are not yet live in the West

Nintendo has introduced new guidelines for the Nintendo Switch 2 eShop in Asia to seemingly combat low-quality games.

Back in May, the company updated the Nintendo Switch eShop to filter out cheap games and "slop," and now, it has implemented further improvements by releasing new guidelines in Japan and some other Asian countries.

As IGN reports, the guidelines target several topics, including game bundles, and how they can be sold, sensitive content restrictions, prohibitions on inaccurate product descriptions, and when and how product information can be updated.

Firstly, in the first year of a game's release, only a maximum of five game bundles may be distributed. The number can then increase for each year the game is available, up to a maximum of eight different bundles.

This new restriction appears to be a way to combat how publishers will constantly push bundles on the store to keep their games in the eShop top charts.

Nintendo is also tackling sensitive content on the platform, which includes "sexualization of children, overly sexual content, discrimination and hate, exploitation of social issues, instructing criminal activity, and political statements". Inaccurate descriptions will now be forbidden.

"It is prohibited to provide inaccurate descriptions of the contents of a product. It is prohibited to provide description of the content of a product as under development if it is not expected to be implemented in the product," the guidelines read.

Finally, publishers and developers will no longer be able to alter their game descriptions without good cause and are now prohibited from changing information on a game's product page after it has gone live.

Developers will also need to contact Nintendo representatives if they intend to distribute an application "that does not include game elements."

These new guidelines are not yet live in the West, but we expect something similar soon.

• The Nintendo Switch 2 is the company’s least ambitious console to date, but its improvements are astronomical
• I’ve spent 40 hours exploring Death Stranding 2: On the Beach, and it’s an incredible sequel that builds upon its unique predecessor to become a masterpiece
• I’ve spent 150 hours with The Legend of Zelda: Breath of the Wild, and the Switch 2 Edition is an incredible upgrade