Networking

Back in my childhood hometown in Calhoun County, West Virginia, my high school friend Bill Bailey says his DSL connection gives him an “average speed of 27Kbps.” That’s Kilobits per second. He might do better with a modem. It’s not just him. Tens of millions of Americans who live in the country can only dream of having broadband. The answer to their internet dreams may lie in the skies above them, with SpaceX Starlink satellites.

SpaceX has applied for the Federal Communication Commission’s up-to $16 billion Rural Digital Opportunity Fund (RDOF). This is a plan to bring broadband — with download speeds of at least 25Mbps — to six million homes and businesses, which currently have no broadband. As part of its presentation, Starlink showed internet performance tests with download speeds of between 102Mbps to 103Mbps, upload speeds of 40.5Mbps to not quite 42Mbps, and a latency of 18 milliseconds to 19 milliseconds. That’s much better than conventional satellite internet, comparable to low-end cable internet, and far beyond what most rural internet users can get.
Other independent third-party tests are showing lower performance numbers. Users posting to TestMy.Net are showing an average download speed of 37.04Mbps, with a top speed of 91.04Mbps. Other tests show a top download number of 103Mbps, an upload speed of 41.99Mbps, and a latency of 18 milliseconds. That’s still much better broadband than many rural users have been seeing.
Of course, the Linux-powered Starlink satellites are still in beta. And, with about 775 Starlink satellites now in orbit, the service is far short of its initial goal of 12,000 satellites. SpaceX has applied to the FCC to launch 30,000 Starlink satellites. According to SpaceX founder and CEO Elon Musk, SpaceX needs about 400 Starlink satellites to provide “minor” coverage and 800 for “moderate” coverage.
Pent-up demand for Starlink’s fast Low-Earth Orbit (LEO) internet is also increasing. SpaceX recently applied for an FCC license to roll out five million ‘UFO on a stick’ end-user terminals over its original request for a million terminals. This came after 700,000 US residents signed up to be updated about the service’s availability. 
It’s not just broadband-hungry individuals who want Starlink services. Rural governments, such as the Federation of Northern Ontario Municipalities (FONOM), are also looking to the sky for broadband. “We know today our citizens require greater connectivity than 50/10 megabits per second,” said FONOM president Danny Whalen in a recent press release. And, “FONOM believes that the Starlink program is our best option.”
So, why aren’t we seeing more beta testers or even an early release program? The answer is there aren’t enough terminals in the production pipeline. A close reading of the SpaceX FCC request to modify the Starlink satellite constellation orbits reveals SpaceX is “on track to produce thousands of consumer user terminals per month, heading toward high-rate production.” If they’re on track to produce thousands, that implies they’re now only producing hundreds of terminals per month. 
This theory is further supported by a search for SpaceX Starlink job openings. This LinkedIn search revealed that the company is actively looking for “talented production associates to help establish a new manufacturing and test processes and eventually ramp to full-scale production for the user terminal.” The company is also looking “for skilled and well-rounded maintenance technicians to help ensure high productivity and reliability of our high volume Starlink User Terminal manufacturing facility.” SpaceX is also seeking “manufacturing engineers to spearhead the development of millions of consumer-facing devices that will sit in our customers’ homes.”
Put it all together and what I see is that, while Starlink satellites are capable of delivering the broadband goods, it will still be months more before enough SpaceX’s CA-based factory can meet the demand for first hundreds of thousands and eventually millions of terminals. 
Related Stories: More

Image: Supplied
Ericsson announced on Friday that it has acquired Cradlepoint in a deal that values the enterprise at $1.1 billion.
The Swedish carrier equipment manufacturer said Cradlepoint had SEK1.2 billion in sales for 2019, which is approximately $137 million, with a margin of 61%.
After dragging Ericsson’s margins down by 1% in 2021 and 2022, it is expected the deal will start to make a contribution to operating cash-flow in 2022.
Cradlepoint and its over 650 staff will be a standalone subsidiary of Ericsson, and remain headquartered in Boise, Idaho with offices in Silicon Valley, the UK, and Australia. It will be part of Ericsson’s business area technologies and new businesses segment.
The company describes itself as providing “cloud-delivered wireless edge solutions for branch, mobile, and IoT networks”. Cradlepoint has over 20,000 customers, 1,500 channel partners, and over 1 million SaaS subscriptions.
“Ericsson is uniquely positioned to build on Cradlepoint’s leadership position in wireless edge and the wireless WAN market,” CEO Börje Ekholm said.
“Combining the scale of our market access and established relationships with the world’s biggest mobile operators we are making a strong investment to support our customers to grow in this exciting market.”
In a call, Ekholm said Cradlepoint is currently focused on the US market, but that the subsidiary has significant potential outside America.
On the other side of the planet, and the weekend, Indian giant HCL entered into a scheme on Monday to acquire the shares of Australian IT services vendor DWS for a valuation of AU$162 million. This deal will also be cash funded, and should the deal get the proper approvals, it will be implemented on December 8.
At the end of June, DWS had 770 full time equivalent staff.
For its full year results posted in August, DWS reported a 2.7% increase in revenue to AU$168 million, with reported EBITDA down AU$1.45 million to AU$20.5 million, and net profit down by AU$2.8 million to AU$7.5 million.
DWS said COVID-19 had an impact on its results, with its Symplicit business receiving AU$400,000 in JobKeeper payments and applying for rent relief on its Sydney and Melbourne offices.
“DWS has grown revenue and underlying EBITDA with growth in federal government and state government client work greater than reductions in banking and finance and IT&C client work,” it said.
Related Coverage More

The default login option for agents used by the Australian Taxation Office (ATO) is vulnerable to a code replay attack, security researchers Ben Frengley and Vanessa Teague said.
Writing in a blog post, the pair described that an attacker could use a malicious login form to capture user details, which the attacker could then use to login into other accounts held by the myGovID user.
The nub of the attack is that when a myGovID user attempts to login into a site, they are asked to input a four-digit code into the myGovID smartphone app to verify the login — no passwords are used, and the only identifying piece of information is an email address.
If the attacker can capture an email address, that can be used by the attacker to log into another myGovID service and replay the generated code to the user to enter into the myGovID app. Once the code is entered, the user will believe they are logged into a proper site, while the attacker can simultaneously log into their account elsewhere.
The user is not alerted to the other login taking place.
[embedded content]
“This attack is detectable by a diligent user who understands the protocol well enough to know that they should only accept 4-digit codes from mygovid.gov.au (and knows how to check for TLS),” the pair wrote.
“However we believe that there are very few users in this category, because it is a counter-intuitive protocol designed to reverse the information flow relative to what users are accustomed to.”
The suggested short term mitigation from the researchers is to inform users about what site is requesting a login, and for the long term, the pair recommended ditching the framework altogether.
“In the long run, the [Trusted Digital Identity Framework] and all its current implementations should be deprecated and replaced with an open standard such as OpenID Connect or a protocol modelled on that of a nation with an existing secure public key infrastructure such as Belgium or Estonia,” they wrote.
“The implementation and design documentation should be openly available to the Australian public to allow for the identification and responsible disclosure of other vulnerabilities.
“We have no reason to believe that this is the only, or the worst, vulnerability in this system. Its complex nature and the desire to hide information makes enforcing and validating correct, secure behaviour close to impossible.”
For users, the pair recommended they do not use myGovID unless unavoidable, and in that case, to ensure they only receive codes from the mygovid.gov.au site.
“This unlikely to work in practice for most users, who will struggle to recognise a secure website with the right URL,” they said.
The pair said they informed the Australian Signals Directorate of the issue on August 19, and were told on Friday by the ATO that “they did not intend to change the protocol, at which point we immediately informed them that we would make a warning to users public”.
In October, the Digital Transformation Agency said almost 7,000 Australians had created a myGovID.
Also on Monday morning, the ATO announced it has signed a three-year, AU$11.4 million deal with Vocus for managed network services.
“The contract will see Vocus provide up to 230 services across 80 ATO sites, on its fully separated secure network,” Vocus said.
“The types of services include IP WAN, internet and data centre connectivity for all existing and future ATO sites.”
The contract has three potential two-year extensions.
Related Coverage More

Image: NBN
Aussie Broadband has seen strong demand with around 5.5% of its customer base jumping onto the 1Gbps plans the ISP began to offer in May.
“In our view, much stronger [demand] than we actually anticipated in those plans,” Britt told the Joint Standing Committee on the National Broadband Network on Friday morning.
“We think it could get to 10% within a couple of years.”
As it currently stands though, NBN is only offering its 500-1000Mbps plans on 18% of its network, which consists of all its fibre to the premises footprint, and an initial 7% of the HFC network.
“We can provide it on some parts of the HFC network, it’s roughly 9% of the HFC footprint today, and NBN has committed to increasing that footprint further,” Britt said.
“We believe in time, it might be available to the fibre to the curb footprint as well, the technology certainly supports it, it’s something that NBN needs to enable.
“So if that all flowed through, then probably about 50% of the footprint in total would be able to get those sorts of speeds.”
Britt said there was pent-up demand in HFC areas for gigabit speeds, but users were waiting for a capable connection.
The Aussie Broadband chief added that there are “very few people that actually need that kind of speed”.
“It’s one of those things. If it’s available, and I’ve got the income to support it — because I mean that ultrafast plan is AU$150 a month sort of thing — then I’ll want it and I have it. And when they do an update, they might use it and they’ll use it for like a download [that] might take 10 minutes or something like that, instead of taking two hours on another type of service, and they want that,” he said.
“I guess I’m in that category as well; I traditionally live in a fixed wireless area, but I have a link to a neighbouring estate that has a mate that’s got fibre to the premises, and so I get my ultrafast via an extension.”
At the other end of the market, Britt said he could definitely see 5G starting to eat the lower end.
“I absolutely do think that 5G will have an impact at the lower end of the market, so people with a lower usage, may be less than 100 gig … but I think at the more family to high end of the market parties, 5G is not really a solution that will play out in that regard,” he said.
Returning to a familiar complaint in the form of the CVC, Britt once again called for NBN to charge only for access, not capacity.
“I guess the artificial constraints around CVC were probably relevant 10 years ago when it was dreamed up, but I think it’s time we moved on from that,” he said.
On the topic of NBN in the enterprise market, Aussie Broadband is not as critical as fellow telco Vocus has previously been.
“The thing to consider is that NBN has an obligation to provide basic level services to that building, and that building might have lots of large businesses, but it might also have the coffee shop, and the Subway, and whatever else is there, NBN has to provide a basic TC4 service to that,” he said.
“For me, if you’re going to have to service those small businesses anyway, then you might as well run the fibre in, and if that opens up enterprise opportunities, well that’s a nice bonus.
“I don’t see that it’s a misuse of taxpayers money like what Vocus does. I see it that they have to provide the service to others in the building anyway, and then if I’ve put up in a seventh option in the building, well, then that’s a bonus.”
At the start of the year, NBN said it would retreat from the practice of directly signing up enterprise customers, following criticism from the likes of Vocus, and would instead need retailers to have a direct relationship with enterprises.
However, Britt sees a third way involving a panel of retailers.
“I think where [NBN] operates today is probably about right. Where it was operating six months ago, perhaps was going too far in terms of constructing deals and things like that,” the Aussie Broadband chief said.
“I think where we would like to probably see it get to is obviously some very, very large customers want to deal directly with NBN, but I think it probably needs to move to more [of] an arrangement where if these large providers do want to construct a deal with NBN but then there’s a panel of providers that can work with them, and essentially tender for the work, as opposed to every provider trying to construct the direct deal with that company separately.”
Aussie Broadband is currently in the process of heading towards its AU$40 million initial public offering (IPO), and earlier in the week offered its customers a chance to purchase between 2,000 and 10,000 shares each, valued at AU$1 per share, for a potential total customer stake of AU$10 million.
The company is selling between 30 million to 40 million shares in the IPO, which would give the ISP market capitalisation of between AU$180 million to AU$190 million, and an enterprise value of approximately AU$150 million. Following the IPO, the current shareholders in the company would have a total stake of around 78.7% to 83.1% depending on how many shares are sold.
In fiscal terms, Aussie Broadband has been able to grow its revenue from AU$49.3 million in FY18 to AU$190.5 million in FY20, and forecasts it will record AU$338 million in the coming year. In statutory terms, the company has not posted a net profit since at least FY17.
The company said it did not intend to pay dividends to shareholders, and would instead be investing the money back into the business. Of the money raised, between AU$20 million to AU$26.5 million would be used to build out Aussie Broadband’s own fibre optic backhaul network, AU$7 million to AU$10 million will be set aside for working capital, and around AU$3 million to AU$3.5 million will be used on costs associated with the IPO.
Related Coverage More

Qualcomm, Ericsson, and US Cellular have revealed what is thought to be the first extended-range 5G NR mmWave data call.

Performed on a commercial network, the call paves the way for 5G applications in rural areas for both businesses and consumers by extending the service possibilities of Fixed Wireless Access (FWA) installations.
Taking place over a 5km distance in Janesville, Wisconsin, the call was made by applying extended-range software to Ericsson hardware, including the AIR5121 and Baseband 6630, together with 5G-enabled customer premises equipment containing a Qualcomm X55 5G modem and a QTM527 mmWave antenna module.
The companies were able to establish speeds of over 100 Mbps. According to Qualcomm, the “milestone” call will help “bridge the digital divide” — as it shows that 5G can enable rapid speeds across areas traditionally neglected when it comes to broadband, including rural and remote locations. 
See also: Qualcomm brings 5G to Always On, Always Connected PCs with debut of Snapdragon 8cx Gen 2
At a time when many of us are working from home and organizations ranging from healthcare providers to schools have been forced to rely on internet connectivity to operate, FWA utilizing 5G mmWave technology could solve connectivity issues faced by consumers and businesses alike. 
“This achievement redefines the perception of 5G mmWave spectrum as an urban- or high-density-only deployment technology and offers new opportunities to use current infrastructure for broader 5G coverage,” Qualcomm says. 
The company added that the milestone also opens the door to new FWA use cases, such as tackling the “last mile.” This issue, faced by Internet Service Providers (ISPs) worldwide, relates to the slowdown of internet connections from central connectivity points to consumer devices due to location and traffic, which can result in performance downgrades experienced by customers. 
CNET: Galaxy Z Fold 2: A drool-worthy foldable that fixes the original’s disastrous design
“This major milestone of using mmWave for an extended-range 5G data transfer is paving the way to implement fixed broadband services for broad coverage in urban, suburban, and rural environments,” commented Alejandro Holcman, senior vice president of engineering at Qualcomm. “With the introduction of the Qualcomm QTM527 mmWave antenna module as part of the Qualcomm Snapdragon X55 5G Modem-RF System, we are empowering operators and OEMs to offer high-performance, extended-range multi-gigabit 5G broadband to their customers.”
TechRepublic: Verizon pairing Real Time Kinematics with 5G for precise location services, road safety, and IoT
The US chipmaker has poured millions of dollars into the research and development of 5G, and in recent months, has released 5G-enabled processor designs suitable for smartphones and laptops.
The Snapdragon 865 and 865 Plus chips are aimed at premium 5G smartphones, whereas the Snapdragon 690 has been designed for mid-range devices. Earlier this month, Qualcomm debuted the Snapdragon 8cx Gen 2, a 5G processor for Always On, Always Connected PCs. 
According to Gartner, worldwide 5G network infrastructure market revenue will reach $8.1 billion, despite the disruption caused by COVID-19. 
Previous and related coverage
Have a tip? Get in touch securely via WhatsApp | Signal at +447713 025 499, or over at Keybase: charlie0 More

Victorian-based ISP Aussie Broadband is giving any customers it had at September 7 a chance to purchase between 2,000 and 10,000 shares as part of its initial public offering (IPO).
“This is an unusual offer for a company to make, but it aligns with our value of being good to people,” Aussie Broadband managing director Phil Britt said in an email to customers.
“In the spirit of one of our other values, no bullsh*t, the allocation of shares will be determined on a first come, first served basis if demand exceeds supply.”
The company is selling between 30 million to 40 million shares at a value of AU$1 each, which would give the ISP market capitalisation of between AU$180 million to AU$190 million, and an enterprise value of approximately AU$150 million. Following the IPO, the current shareholders in the company would have a total stake of around 78.7% to 83.1% depending on how many shares are sold.
In its prospectus, the company said it did not intend to pay dividends to shareholders, and would instead be investing the money back into the business. Of the money raised, between AU$20 million to AU$26.5 million would be used to build out Aussie Broadband’s own fibre optic backhaul network, AU$7 million to AU$10 million will be set aside for working capital, and around AU$3 million to AU$3.5 million will be used on costs associated with the IPO.
Aussie Broadband currently has wholesale agreements with Telstra, Vocus, and Opticomm, as well as Symbio for VoIP services. It expects the build out of its fibre network will replace up to 63% of its current leased backhaul to NBN points of interconnect, and enable it to end dark fibre leases between capital city data centres.
“Capital expenditure on this project is budgeted to be AU$67 million and is expected to require two years to complete. The first two stages of deployment of the company’s fibre project commenced in May 2020 and are expected to be completed during September 2020,” it said in its prospectus
“The company estimates that the fibre project will have a minimum useful life of 25 years, and will improve the company’s ability to provide redundancy and increase capacity loads to meet market demand.”
In fiscal terms, Aussie Broadband has been able to grow its revenue from AU$49.3 million in FY18 to AU$190.5 million in FY20, and forecasts it will record AU$338 million in the coming year.
Statutory earnings before interest, tax, depreciation, and amortisation (EBITDA) was reported as AU$3 million for the 2020 fiscal year, and forecast to be AU$10.8 million for 2021.
However, depreciation and financing expenses resulted in the company posting a statutory loss of AU$12.3 million, with that number to widen to AU$15.2 million next year. In statutory terms, the company has not posted a net profit since at least FY17.
The company said it has funded its operations thus far from business equity and debt capital raising, and it would continue to do so in future.
Shares in Aussie Broadband are expected to start trading on the ASX on October 27.
Related Coverage More

Ever since day one, people have thought Chromebooks just ran the Chrome browser and that was it. Actually, it was always more complicated than that. Underneath that Chrome browser was a thin Linux distribution, Chrome OS. Now, Kent Duke, a writer and hardcore Chrome OS fan, has found that Google is teasing apart the browser and the operating system into two separate entries.

Why does this matter to anyone except the most dedicated Chromebook nerd? You see, today when you buy a Chromebook, it comes with an end-of-life date, its Auto Update Expiration (AUE) date. This is a generous six and a half years after any specific Chromebook is released. You can find yours by going to the Update Schedule section in your machine’s Chrome OS settings. You’ll find it underneath the About Chrome OS section, under the Additional details menu. 
Six and a half years is a lot, but Chromebooks, thanks to their old solid-state build, can last much longer. For example, the only reason I’m no longer using the very first commercial Chromebook, the Samsung Series 5, is I sat on it by accident. Whoops! 
That wasn’t very security conscious of me. But now, as Duke pointed out:

Currently, Chrome is intertwined deeply with Chrome OS, meaning Google has to compile and ship one monolithic package to the update channels. While that isn’t an issue in itself, the major problem lies when a Chromebook hits AUE, or end of life. Just like on an Android phone, when your Chromebook hits AUE, you lose out on new Chrome OS updates. Losing out on a Chrome OS update also means that Chrome itself won’t get updated either, which leaves the browser outdated, vulnerable, and unable to take advantage of updated platforms on the web.

 Duke continued, “Since this Chrome binary is distributed separately from Chrome OS, Google can easily update the Chrome binary independently from the operating system. That means even if your Chromebook hits AUE, your browser will at least get the latest and greatest features — and critically, security fixes — from Google.” Costs conscious schools in particular, would be able to keep using older Chromebooks without any significant security worries.
Yes, the Linux underneath it might still have security problems. But, despite the security hype, there have never been any truly serious desktop Linux security problems. This is especially true of Chromebooks, where their underlying Linux foundation has a tiny attack surface. The same can’t be said for the browser, which is under constant attack. 
Here’s how it works. This new alpha approach is called Linux And Chrome OS, or Lacros. Today, the Chrome web browser binary is mixed together with Chrome OS System UI and Linux. WIth Lacros the existing Chrome web browser Chrome OS is renamed ash-chrome. To this, Google’s developers have added the Linux Chrome Web browser and renamed it to lacros-chrome. Yes, this means you can run two versions of Chrome, the one incorporated with Chrome OS and the Linux version on the same machine. 
To improve the Linux browser’s lacros-chrome performance, the programmers have also improved Chrome OS’s Wayland display server protocol. This lets you run two versions of the Chrome web browser without significant performance costs. Lacros uses ozone as an abstraction layer for graphics and event handling. Ozone, in turn, has a “backend” with client-side support for the Wayland compositor protocol.
The application programming interface (API) boundary between the two browsers will be semi-stable. For now, you can run two browsers so long as they’re only up to two versions apart. So, for example, you can run Chrome OS 78 with its built-in Chrome web browser version 78 and the Chrome web browser version 80. In the future, you’ll be able to run a Chrome web browser that’s much newer than the version embedded within an older edition of Chrome OS. 
What all that means for you is that by this time in 2021, you may be able to run your old Chromebook safely with the newest version of Chrome. Besides just being handy to you personally, it may also save your school system and business a good deal of money too.
Related Stories: More

Nokia on Tuesday rolled out enhancements to its Self-Organizing Network software, bringing more automation to network operations. The vendor-agnostic software should help communication service providers more easily manage complex 5G networks, Nokia says. 

Prior versions of the software automated some time-consuming tasks for network operators, “but you still needed individuals to decide which part of the networks needed to be optimized,” Brian McCann, Chief Product Officer for Nokia Software, said to ZDNet, “to identify a problem, to identify the right solution and verify that it fixed the problem and monitor the outcome.”
Prior versions of the software automated some time-consuming tasks for network operators, “but you still needed individuals to decide which part of the networks needed to be optimized,” Brian McCann, Chief Product Officer for Nokia Software, said to ZDNet, “to identify a problem, to identify the right solution and verify that it fixed the problem and monitor the outcome.”
The latest release uses machine learning to detect network issues and self-correct any deviations based on a set of defined objectives, such as latency or throughput levels. It can also leverage insights to further improve the solution itself. 
That should help as 5G deployments increase, McCann said. 
“There’s a lot of opportunity with 5G, but it’s a much of a real-time environment, so the reduction and almost elimination of human interaction is really a requirement,” he said. 
The software has a vendor-agnostic network slice management function to automate the radio slice life cycle and resource optimization. Its machine learning capabilities will allow it to optimize each slice separately to ensure better network availability and quality. More