Does anyone know when starlink is going to be available for RVs or campers? I'd love to bring the internet while I camp around California, but right now there are range restrictions.
It’s in beta now. Set mine up. If it works the way it does now when the masses join you are going to really like it. Anecdotes in another comment here.
Indeed there are range restrictions with Starlink. What I did was order Starlink to the nearest post office to a place that has a unique intersection of lovely camping grounds that doesn't have internet access. The idea is to rotate between them. Use https://sebsebmc.github.io/starlink-coverage/index.html to calculate how big your cell is and then intersect with your camping guide :)
Hi, yeah I created that and haven't updated it in a long time. You can pull the github repo and run it yourself to get a new map but there are some issues that have cropped up since that have prevented me from being able to update the site myself.
I did manage to run the simulation in May and surprisingly the data appeared to indicate less stars but higher average coverage time (by a few minutes). I didn't push this update because it wasn't much of a change from the current map, but I probably should have since that is interesting in itself.
I will note that the cells that I have rendering on the site there are for performance and download size reasons, and were never meant to correspond to any geographical restrictions that Starlink has put in place.
Unless you have huge batteries, 90w constant draw is 2.1kW a day.
With all that other hardware, you'd have to be essentially tethered to the grid, or be recharging your batteries from grid regularly. (looks like you have 6x100 watt solar panels which won't do much for that kind of consumption)
I just stumbled onto Speedify a few minutes before I noticed this article surprisingly. I wanted to try it but unfortunately it doesn't take crypto and has no free trial
That's a shame. They used to have a free trial with a data cap which I used intermittently while traveling for years before converting to a paying customer a year ago.
Maybe they looked at their numbers amd couldn't didn't see the free tier paying for itself in conversions, but it did convert me.
We still have the 2GB per month free tier with no registration. It's available everywhere but Linux... unfortunately, without us doing anything really creepy (which we won't do), it was just too common for Linux users to trick us into endlessly starting the 2GB over. I think there was a script for automating it making the rounds on some sites.
AFAIK multi-path is not too common, though most people who do multipath are using MPTCP.
Last I checked, It's part of LTE standard, but used exclusively in South Korea. Some providers across the world provide multi-path gateways: in France, OVH has OverTheBox which is line-provider agnostic, Free has ADSL+4G aggregation where fiber isn't available. In Switzerland, Swisscom also has some multi-path offer.
Also, Apple's Siri uses MPTCP. (And developers love the fact that only Siri can use it, and not other apps).
Google/Android said no to MPTCP, saying QUIC/HTTP3 already has the provisions needed to do multipath, but I don't know if Google actually uses it somewhere.
From experience, Free's ADSL+4G aggregation is stupid. The router needs ADSL to be up all the time and aggregation is unpredictable and opaque. Being the carrier for both ADSL and 4G would've enabled them to do multi-path magic and assign the same IP to both endpoints yet they don't. It seems to be basic bonding.
QUIC has some basic features for migrating a session to another link, but it doesn't support multipath in that sense. Multipath QUIC (MPQUIC) is a hot research area right now, and it's being standardized.
MPTCP's design was fundamentally incompatible with QUIC's UDP transport and session model.
From what I understand, neither MPTCP nor MPQUIC are designed to share multipath routing via a gateway (e.g. a bunch of devices sharing a multipath aggregation of 3 4G modems). It's all designed for a single device with multiple interfaces.
I was wondering about this exact same thing the other day. Working from home makes my mobile data go wasted every month. I actually have a company phone in addition to a personal one.
Wow. I am in the middle of a big city and I have 50Mbps down and 11 down and you have that speed in the middle of nowhere... the future sounds amazing.
Tell me about it. My 1Gbps docsis3 has an upload of 35Mpbs because of 1980s cable tv frequency choice reasons. I’m slightly impressed with StarLink and slightly annoyed at my main ISP.
This is on the edge of nowhere, but still cool. The only other option I had here was dialup or if I wanted to pay outrageous rates, I could get a .5-1Mbps T1 only because there is a ranger station not too far off that put one in and I can access some of the trunk hardware a few miles away.
latency will go still down significantly as v2 and v3 join the constellation with laser interlinks. For long connections, they will be theoretically faster than fiber due to geographical constraints on straight lines and routing.
Side notes: The more I thought about it. In case anyone is interested:
- The setup is stupid easy. Place dish outside, plug in. That's about all.
- It's weird having a dish that points straight up, used to them all pointing south (hemisphere dependent of course)
- The router box (thing?) is a bad design. Stupid little triangle that tips over all the time. Seems cheap.
- Shipped with no wifi encryption. I thought that was weird. I bet this changes to for real release.
- Got a few "outside your home area" messages without moving. Errors.
- It is good for streaming or websites, but currently I have it on the ground and it seems like it loses connection fairly often. I'll put in up on a poll now that I know...
- The dish is heated. GOOD. I was worried about this with the amount of snow it's going to get.
I don't really geek out over these things. It's good. It's exactly what it should be I think. We'll see how it lasts the winter. I expect the reception to get better when it's mounted, but it'll be interesting to see how white-out snow effects the performance.
Got a source for that? When I last looked into it, their site said it was a technical limitation and that there were no plans for creating a truly mobile receiver.
Eagle-eyed observers have noticed Starlink antennas on SpaceX's drone ships. So mobile capability is definitely already implemented, it's just not publicly available yet.
(Edit: And evidently some regular users are already beta-testing it too.)
Given the size of cutlery, coupled with the fact that multiple bowls will fit inside eachother, is this overcompensation for an extreme maximalist lifestyle on the electronics side?
And soon 7 (!!!) internet links! Anyway, I just find it interesting how individuals (including myself) describe themselves and rationalize their descriptions.
It sounds like OP has found that six different ways to connect to the Internet is the minimum that is viable for him, and that meeting that is important enough to justify spending a large portion of his limited budget (both financially and in terms of space in/on his vehicle).
If one fork does the job, one fork is enough. If five way to connect aren’t enough, then maybe six will be enough :)
I disagree. The highly specialized monitoring equipment and dual high gains antennas and the plethora of other bits and pieces are NOT required to run a "a meeting" or a "netflix stream". A phone with a sim card usually suffice. All this frivolous spending points out that this is a hobby. Nothing wrong with having hobbies, of course, but the spending is not compatible with the definition of minimalism.
A minimalist online meeting is an IRC connection over 3G.
You're underestimating how terrible internet connectivity is in rural Australia. A lot of places will only have 1 provider that offers service, if any at all.
Admittedly a low-effort post, and too late to delete, but just to explain, this whole idea of having a single bowl (used to collect alms) is a thing in monastic/ascetic traditions. Which I always get the sense minimalism is vaguely inspired by. Anyway, whatever, it was off-topic.
I did actually prototype multi-link PPP over multiple GRE tunnels as a way of bonding multiple links, with (some level of) redundancy, with the intention of offering that as a product at the ISP I worked. The idea was, basically "up to four ADSL links, possibly from multiple providers, each used as a member of the MPPP bundle, then dial-up of some sort as last resort".
Also tested ADSL as a bond/fallback for a leased line, around then. Ah, the days...
It is for my family. It may not be for yours. When they reach the age where they have their own cellphones / social media media then it's different. Freshly popped popcorn from the microwave + salt + butter and introducing kiddos to "the dish" movie after _they have been to the actual dish_ is A+ and highly recommend.
Yes, meant to add that if folks are misunderstanding, perhaps you could clarify. (But I was not yet awake.)
From the text however, its not clear how often the kids are around. Maybe they could benefit from some downtime, maybe not. Maybe it is their link to school and would therefore be even... crucial. In absence of info people tend to speculate. ;-)
Also, I have a stash of important movies that I keep for posterity. Even at a normal home with reliable internet. I don't want to pay for rental every single time, or have them go off streaming at random as they so often do. You'd be surprised what can fit on an SD card or micro usb flash these days.
Deliberately so, this information is private and personal. Especially right now due to lockdowns in Australia. I have publically shared this recently because enough is enough:
Sorry, not trying to pierce your privacy, it was merely a factor on the discussion. Also, I don't use twitter for various reasons, so couldn't read the blurb there.
Enough to take one DVD with you plus a USB based drive. It's not that much space. And depending on your jurisdiction, it might even be legal to rip a DVD and take the ripped version with you for the trip on say a tablet's flash storage. No need for the internet.
I don't have young kids myself, but I feel like if you are camping for long enough(like if you live in a van), it becomes less of a "escape the modern comforts to reset" kind of deal, and you'll want to start blending in modern parts of life.
Young kids get over poor or no internet, but it’s not instantaneous. It’s part of that whole actually being a child thing rather than simply childish behavior.
Can confirm. At 8pm every evening a cron job adds a few firewall rules that block a few devices in the house. Much better than arguing w/ the little one about stopping what they were doing and getting ready for bed.
OP lives in Australia. Which means that functioning telecommunication is an absolute must given the number of living lethal things that are out to kill you.
For what it's worth it's interesting (but possibly explained by the ridiculous cost of anything not SpaceX) OP doesn't have a satellite internet uplink too.
Nothing is out to kill you. It might do it by accident, or if you startle it by trying to use a toilet or wear your shoes, but there’s no malice in it.
My "camping" is "your living" but in nature instead of being a slave to a circa $912,382 home loan in Sydney, Australia. When I want to go "your camping" I house-sit or stay at friends houses.
On my reading I thought - no need to justify this obviously cool project! I think sometimes people feel the need to have a good reason to do something, even when they are doing it because it's just intrinsically awesome. I've totally overengineered my home network, and have spent more than I probably should have, but I don't think it's a waste at all because there is something deeply satisfying about it. But when I defend my purchases to others I am tempted to say, "I need good internet for work", or "we run many IoT devices" and those things are true, but could have been completed with much less time, money and effort.
It's a learning experience, and a hobby. And there's absolutely nothing wrong with finding joy in that, and absolutely no reason to have to explain it to others, because you're likely the only person who gets it.
For sure. We doing a camping family reunion every year at a site with no service and it's great. Camping for me is about particular and very useful kind of giving up. We've done things at that reunion together that never would have happened if kids (or adults) could wander off with their devices. E.g., one year I brought many cheap copies of "A Midsummer Night's Dream" and over two nights we did a firelight readthrough. Many of the kids were suspicious at first and then really got into it.
I remember when HN used to discuss cool technology hacks and not incessant hand-wringing about the plague of internet-connected children at campgrounds or the deluge of fraudulent handwriting in the postal system.
I would then argue that it's that very "stick our heads in the sand and pretend nothing's wrong" attitude that contributed to the massive user-hostile mess that tech, broadly, has become.
Well, I think he's mostly doing that, because he is actually working remotely, while camping:
> If you are going to work remotely anywhere in Australia from a van, you need damn good internet.
In that context it may not be an overkill. If you're 2-3 days away from nearest wifi enabled spot, then having a solid backup strategy might be a good idea...
Still using a Alfa AWUS036ACH for a wireless connection: 802.11ac wave1 from 2015.
There's simply nothing better available, in 6 years of intervening time.
This world would be so so much more awesome if there were some really really good USB wifi adapters. This is yet another further leap, but I also think of all the guides to turning the RPi into an access point, via usb adapters: it starts out working fairly well, but the hardware either has hard maximum client limits or just fails to work well once you get to half a dozen devices. Even if you have something like PCIe or M.2 expandability, options are still extremely limited, & availability is incredibly poor.
I'm a huge fan of Alfa gear, for a long time. They do an amazingly good job.
AWUS036ACH is one of Alfa's most up to date USB cards. It is, however based off an RTL8812AU chipset[1], which is 8 years old now[2]. I'd love at least to have a good client chipset & equipment that supports WiFi6 that has good RF performance. I still think the market is overly structured & segmented. No one makes AP-class chips for USB, but to be honest I really think they could & should & that it'd be exceptionally useful & good. But no one dares to threaten the AP market like that.
Thankfully in wifi, RF engineering & antennas count for a huge amount. You still have great gear. But it does hurt me to see what seems like rampant tech-stagnation & neglect, of such a vital, interesting, basic part of modern communication technology: our wireless systems.
This is a pretty impressive setup. Recently I lived in an RV for a year and my setup was much more janky than this...
Basically consisted of a gl.inet router that I plugged in a old Moto X4. This allowed me to get wifi from whichever campsite we stayed at (always garbage) and then I would pay for 100GB from Cricket and T-Mobile. I would swap sims when I would run out of data about half way through the month.
Then I discovered I could get unlimited data on Cricket with a regular phone plan if I changed the TTL on the router. Ended up being the most stable option for the last few months.
I setup a firewall rule in openwrt that said anything coming over USB needed to set the TTL to 65 so it looked like the data was coming from the phone itself and not from the tether.
IIRC, on the plan I was using on Cricket, I would get unlimited data + 15GB of tether. This just byapssed that limit.
I think they can almost figure it out because if I tried to tether without the TTL trick, data wouldn't work. So I'm guessing it still counted the data I was using against some kind of cap?
Years ago, I did this with my jailbroken iphone 3gs. When visiting lots of websites, it would serve up the mobile version, so I assumed that the app was somehow modifying request headers to look like a phone request. I guess not as many sites used TLS back then.
That's the TTL that some carriers use from the phone. So once it hits the router it goes down - meaning the carrier knows you're tethering. Setting it to back to 65 on the router makes it harder for the carrier to detect.
Close, but it's a little bit different than how you describe it. The router is connected between the phone and the other devices in this user's scenario. The default TTL on Linux and Android is 64. If you tether, the packet will go through your phone and have its TTL decremented to 63 (the phone is a gateway). Then when it arrives at the cell tower with a lower TTL, they know you're tethering and drop the packet.
If you set it to 65 on your host device or router, it will be decremented to 64 on the phone - and is now hard to distinguish from real traffic from the phone.
You want it to be one more than the value the phone uses as its default. Then when the packet arrives at the phone via the tether its TTL is decremented and it is passed out to the tower. Tower sees a packet with a TTL that it is expecting and assumes it is phone data.
Same here, tried it on out last time out with the travel trailer. Canceled the plan after we got back home as it was terrible, and not to speak that we were in areas where it should have run flawless. Ended up using my main phone's connection for some of the work i had to do.
Visible's tethering is limited to 5Mbps (according to their pricing page). Have you been getting more than 5Mbps tethering? Do you do anything to get around tethering speed limits?
Deprioritization can be hard in some areas. Verizon has the most customers and the least spectrum at the moment.
I think I've gotten more than 5mbps (or at least was for a time) - I haven't done anything, but I might be wrong about this (I don't rely on tethering most of the time so haven't noticed).
IIRC a long time ago all of Visible's data speed was capped at 5mbps and that was the 'tradeoff' of the low cost service in addition to deprioritization. They eventually dropped that cap, but it's possible it still exists for tethering.
I will plug Starlink for those that can get it though - not mobile yet, but we went from a crappy calnet connection to a stable 50+ sometimes 150+ mbps connection. When they get it working across cells (which Musk tweeted interest about) it'd be a pretty sweet option for an RV. If you're out in the boonies though and looking for stable home internet, it can't be beat (as long as you have a clear view of the sky).
You’d probably like the style of the diagrams created by Julia Evans, then: https://jvns.ca/
I’ve followed her blog for years at this point mostly because I enjoy her diagrams and pretty much always learn something from the way she presents information.
Running all of that separate network equipment seems particularly power hungry when not plugged into a campsite.
I wonder why the author didn't consider a Peplink router which can handle redundant cellular/wifi connections out of the box. That's how I recreate a setup akin to this while I'm on the road in a single device.
Not to say this isn't awesome as is and I'd love an excuse to install a full Ubiquiti stack in my own camper!
Oh, this _seriously_ chews power. I need to do a blogpost up about the eletrical system that makes it all possible. Leave your digits at https://ghuntley.com/newsletter to be notified when it ships. I looked into teltonika (ie https://teltonika-networks.com/product/rut950/ ) but stopped because the modems are Cat4 (similar story with Microtik).
Care to give us a sneak peak - say, what your typical power draw is like? :)
I'm particularly interested in this topic because I've been thinking about power a lot after the Texas blackouts (considering building a bike generator so I can have power in case of a disaster like that one).
Have you considered any of the solar “generators”? Basically they incorporate a solar controller, inverter, and battery into one easy to use package. I typically get at least one outage/year here in the Midwest. I bought a kill-a-watt to measure what my fridge is using (about 2kwh/day) and it looks like one of these devices could keep that running through a longer outage. If you want to go even further you can buy more batteries!
You will likely be disappointed in cellular availability during a long blackout. I'm in the Midwest in a metro area of ~200k. We had a major weather event which took out power for 1-2 weeks. Almost all cellular towers went offline as soon as their batteries died. It seemed there were only 1-2 towers in the entire area that had generator backups.
"The way Speedify used to work is it opened one TCP socket over each Internet connection it used - e.g. one TCP socket over Wi-Fi / cellular / wired Ethernet / etc. With the new protocol, Speedify opens up multiple sockets over each connection (for example 8) - 8 over Wi-Fi / cellular / wired Ethernet / etc."
That seems a bit wasteful towards the server destinations.
The 8 sockets are opened only towards speedify "VPN" server. Once you're out of speedify network, you only have one standard TCP connection, so destination server only sees one connection as expected.
Geoff: some of the photographs have blur spots to conceal something you don't want to be public. Just an FYI that malicious attackers could bypass this blur in some cases by starting with a generated version of the 'before' content and blurring it and iterating different inputs until it matches your blurred photo. This matters for things like a document with an account number or some other secret. It may not matter much for your network adapters (mac addresses, maybe?). Just a heads up.
Thanks wyldfire; the photos that are blurred are the identifiers of my RIPE atlas nodes. It's not a big issue if they become public knowledge. Fully aware of attack vectors related to image bluring.
I think so. If you don’t like the way that looks aesthetically, I’ve overlaid the text with a white background and more text and blurred that, so you get the same blurred effect but you’re blurring nothing important.
Yes, making sure that the file format you picked doesn't support layers (or at least that those layers are flattened) that mistake bit the New York Times a while ago.
You can't reverse a typical blur algorithmically, since it destroys information. You can only guess at the information that was there before the blur, and work from there.
The stronger the blur intensity, the amount of possible starting states to get to that blur increases exponentially. If you blur enough, every pixel is the same color, which obviously has destroyed all information.
But if I know that the blurred content is a social security code written in 12pt Times New Roman - I can perform a blur operation on a million SSNs, and see which one matches most closely to the mess I have on the screen. It's easier with bar codes.
> If you blur enough, every pixel is the same color, which obviously has destroyed all information.
Not true. The ASCII art people have sorted characters by their “shade”; if you know the foreground and background colour of a letter, and you know the single colour of pixel it blurs to, you can still work out what the letter was.
Blurring discards some information and obfuscates other information, but really, given how easy it is to reverse-engineer such things, we should measure the number of bits of information remaining.
If information in image + information about image > bits of information in sensitive data, then in theory you can recover the data.
So simply deleting the data from the image (e.g. blacking it out, removing reflections) is preferable. Saves you a lot of effort!
> if you know the foreground and background colour of a letter, and you know the single colour of pixel it blurs to, you can still work out what the letter was.
Totally agreed, and I like what you said about information remaining - you start to get into fun information theory stuff.
For example, let's say you're given one of the 'shades' you mentioned. I give you #a9a9a9. There's 32 bits of information there. However, if it's a blurred letter of black text on white background, it's always going to be grayscale. There's only 256 possible grayscale values - only 8 bits. Luckily, since there's only 26 lowercase english characters, we can easily fit that into our 255 values. Information is not destroyed! This is the shade->char map you were talking about.
But now what if we blur enough to turn two letters into one shade. That one shade will still be between 0 and 255. But now there's 26*26=676 possibilities that could've created that shade. More inputs than we have outputs for. There's no way to fit more than 256 possible inputs into 8 bits of information. Shade '98' could be 3 or 4 different inputs. However, we can be very clever...
We know that the letter 'u' comes after 'q' almost always. We know that 'x' never appears next to 'j'. We know lots of things, and can supplement the destroyed information, with outside information. We can actually change the whole context for this conversation now. We're not talking about blurring. We're talking about fitting english text into less bits than ought to be possible. We're talking about compression. This is exactly what compression algorithms for english text will do. There's a lot of redundant information in plaintext, just like there's a lot of redundant information in our images of text. An 8x8 pixel character glyph can easily reduce the information to a single pixel. However, there are limits. You can compress english text by 10x, if it's simple enough. You can't compress War and Peace into 5 bytes. You can blur text a few pixels and not destroy information. But you can't blur a paragraph of 12px font by 500px and get your original information.
It’s difficult or impossible to derive the original information from the output, but if you have a small enough keyspace you can generate all of the blurred (“hashed”) versions and compare the results.
To some degree, yes, but that metaphor breaks down because (often/typically) hash functions are designed so that small changes to the input result in large changes to the output.
Anything opaque would work. But aesthetically that's not always preferable. Maybe an opaque central region with blended borders would be a good balance.
I black out all sensitive regions in magenta, blur any bits I suspect might have reflections (blacking them out if I can see that there's actually something there), then copy from elsewhere in the image to get the aesthetics right, then blur to hide the fact I copied stuff.
I have a collapsable ladder which is used to clean the panels. There is absolutely no space on the roof left for Dishy and mounting it up there would cast shadows on the solar system (which is circa 3 * 350w house solar panels). For dishy, it's going to be a matter of running a ethernet cable out the van during the day.
Is it generally accepted that these wavelengths and devices that operate them are safe around humans? Should I feel concerned at all having so many large antennas and amplifiers and what not near me?
I never really stopped to ask this of my phone either, I guess.
This section of the EM spectrum is non-ionizing. So unlike, say, X-rays, you won't get cancer as the waves aren't strong enough to strip an electron off an atom and therefore change your DNA. However, they will heat your cells to some degree (since 2.5GHz is the same freq as a microwave). This can cause cataracts over long exposure to high enough power as your eyes can't regulate heat.
Generally accepted by who? You can find some people who are concerned about this stuff. 5G truthers and all that. But nobody in the evidence-based medicine community.
Nothing is entirely free from risks. That's part of why we do studies on things like this. But, for example, the second study on the page you linked concludes: "At the present state of knowledge there is no positive evidence that pulsed or continuous microwave exposure in the non-thermal range confers elevated risk to the health of the brain."
The mere presence of studies on a subject does not indicate positive evidence of material risk to health.
Alex from Speedify here. Yes, we make a multi-path VPN tunnel, so that calls can be shifted from one link to another without glitches (apps see the same IP address the whole time). To make this work, we do some smart things like retransmit packets that were just sent on the failed link, on the still-working link since they were likely to have been lost.
I can attest that Speedify in redundant mode will handle a link failure in the middle of a Zoom call without a single lost packet or other impact. This isn't failover, though, this is redundancy; it's like having a 1 disk fail in a RAID1.
(I have a similar setup to the OP, but using an array of M.2 4G/5G modem cards connected to external antennas.)
I would recommend EasyTether on an Android phone. It bypasses hotspot specific data limits.
I use a Pi running Easytether connected to a 5G phone, with a travel wifi router connected to the Pi over ethernet. Runs my off-grid internet + security cameras.
How does the "redundant" mode work? You cannot just double TCP packets as within that protocol it's a sign of congestion, it will drastically reduce the throughput. Does this include VPN-ing through some service that can understand duplicates and clean them up?
Maybe it determines whatever route is faster when talking to a specific host for the first time, or every so often, and then sticks to that route for that specific host?
All the connections are likely being used to talk to the same speedify server. Only once the packets make it to that server do they get forwarded to the destination.
As someone mentioned you can round-robin connections, but then you don't get failover (without interruption, anyway).
What I think most of these solutions do is use MPTCP or a custom homebrew protocol to a server that then holds the actual TCP connection you want. So a slightly more complicated VPN. There is an open source project for this:
As someone with a similar setup to the OP, I should mention: OpenMPTCPRouter's networking is completely amateur compared to Speedify. It's a grab bag of half-baked multipath VPN experiments (glorytun over UDP, an outdated MPTCP-based fork of glorytun, and MLVPN), none of which approach the completeness of Spotify's implemention. They generally don't handle changing interface bandwidth, packet reordering, packet aggregation, header compression, mode switching between redundant and aggregation, or any other hard problems here. Hopefully this will save someone the pain of setting these up just to be disappointed.
Also you can simply use multiple uplinks without VPS and it provides better latency. The drawback is that you can't failover/balance protocols that use only one connection. Mosh and browsing work well tho.
On top of this plethora of technology and various providers, doesn't bonding all links to a single solution provider (Speedify) create a single point of failure?
> What kinds of servers do you have?
> We use a combination of dedicated and virtual servers, depending on location and scalability needs. All of our servers have at least 1 Gbps network access and are optimized for bandwidth intensive usage such as live streaming, video conferencing or online gaming.
Let's assume the best case, where Speedify implements this as swarms of redundant servers, so that even if some servers or datacenters are down it may continue to work.
I could not find in the article any mention of disabling Speedify and falling back to using a single provider, or a local bonding solution like described on https://wiki.debian.org/Bonding .
Isn't speedify still a "single point of failure", in the sense that any incident with Speedify (e.g. accounting glitch of any kind) may cause the whole contraption to fall like a house of cards?
The UniFi Dream Machine Pro has a backup WAN link. I plug my backpack nighthawk LTE router into it when it is not in my backpack. If speedify fails, it falls over automatically.
Having a single point of failure here might be OK, if Speedify is so much more reliable than the other solutions (4G wireless in potentially remote areas). Though yeah, I'd like to see an option to failover to a single provider, because it's conceptually so easy to add.
If you're trying to get internet in a van in rural Australia, the chances of Speedify going down are completely insignificant compared to more basic things like not having phone signal.
Alex from Speedify here. Yes, exactly, we have swarms of redundant servers. The servers are very reliable, but in case of failure, you're on another server in 30 seconds.
Your comment makes me realize we may need more options on what to do when the connection fails. Security-types had us add a "kill switch" to make sure traffic couldn't get out when the encrypted tunnel fails. But if you don't care about that, another option to guarantee it would go right out, would make sense too.
That works, and has been done before. It can be confusing to run Speedify and the VPN client on the same box, but if there's a Speedify router box with the internet, and another box with the VPN client, it should just work.
Just wanted to drop in this thread and let you know I think you have an awesome product and I have been happily using speedify for almost a year now. I hope you keep up the great work and are able to keep growing!
It adds a tiny amount of latency, but there are more important factors than latency.
If you are on a Zoom call, tunneling through Speedify in redundant mode will give you the minimum latency of n connections, reduce jitter (latency deviation), and eliminate packet loss.
I’ve looked at Speedify. In fact, I signed up for service with them on my iPad and iPhone as soon as I heard about their service years ago. It doesn’t really get you any higher speed (at least, not on download), what it gets you is a more robust network connection with failover if the primary goes down, although I do see a speed improvement for upload.
On the software-only router side of the house, Speedify at least let’s you tie your multiple links together and get you a more reliable connection. But true aggregation (like LACP) requires that all links be of exactly the same speed — e.g., you do link aggregation on multiple Ethernet ports that are connected to the same switch, and while no one connection will have bandwidth greater than a single Ethernet port, you will have be able to have multiple connections that can all achieve that same level of speed. Their style of bonding isn’t the same thing.
I’ve also looked at OpenMPTCPRouter. On paper, it looks like a better solution to me than Speedify, but in practice it’s not designed to handle multiple different upstream connections with varying levels of performance. I’ve chatted with the author, who lives in France and his ISP connection is 1.25 gigabits per second, so this hasn’t been a priority. He did say that this was going to be something he was going to start working on soon, however.
So, I might be forced to go with something like a six port Protectli pfSense or OPNSense router plus Speedify, at least until I can get something better.
Oh, and Speedify has (or had, maybe they’ve changed this), an effective speed limit that they also implement. I think it’s about 200 mbps. If you’re getting that speed or better over Wi-Fi or Ethernet or whatever your primary network provider is, then they don’t even bother to bring up the other providers.
So, you’re not going to tie multiple upstream sub-1gbps providers together into one giant network connection, because Speedify will just stop bringing up more channels when it thinks you have enough speed.
At least, that was the Behaviour I saw when I was doing my tests. Maybe I had the wrong type of account, so that they were effectively throttling me down?
Hmm. Okay, I did some more testing tonight, and it looks like they do now have more connection modes, including one optimized for streaming, one optimized for raw speed, and one optimized for reliability.
I was able to get some higher speeds with Speedify than I could get with my naked ISP or with NordVPN, and the graphs showed that the cellular 4G connection was actually a significant bonus on top of my DSL.
So, now I guess I’m going to have to build my own Speedify router combined with OPNSense or pfSense, and get those other upstream network connections also set up.
T-Mobile US is using it for their home internet service. With the global supply chain and chips issues, it's been hard for T-Mobile to get enough of them. It seems likely that at this point Nokia might just be doing enterprise sales because of this. If you can't produce enough of them for a customer like T-Mobile US, why would you try selling one to a single end-user? T-Mobile US won't require support, won't be doing returns because they don't like it, won't need an individual shipment, etc. Given that T-Mobile US can't get the supply of these that they need, I'd guess that Nokia is making as many as they can for corporate customers and they don't want to deal with individuals at the moment.
In fact, their website indicates that they're customizing the device with differing specs for different customers with options like an LCD touchscreen and differing antenna designs.
287 comments
[ 0.19 ms ] story [ 179 ms ] threadAs a boat owner... I'm with you!
With all that other hardware, you'd have to be essentially tethered to the grid, or be recharging your batteries from grid regularly. (looks like you have 6x100 watt solar panels which won't do much for that kind of consumption)
https://old.reddit.com/r/Starlink/comments/prgrbb/starlink_r...
Maybe they looked at their numbers amd couldn't didn't see the free tier paying for itself in conversions, but it did convert me.
Last I checked, It's part of LTE standard, but used exclusively in South Korea. Some providers across the world provide multi-path gateways: in France, OVH has OverTheBox which is line-provider agnostic, Free has ADSL+4G aggregation where fiber isn't available. In Switzerland, Swisscom also has some multi-path offer.
Also, Apple's Siri uses MPTCP. (And developers love the fact that only Siri can use it, and not other apps).
Google/Android said no to MPTCP, saying QUIC/HTTP3 already has the provisions needed to do multipath, but I don't know if Google actually uses it somewhere.
Looks like there’s been an API for at least a few years. http://blog.multipath-tcp.org/blog/html/2017/07/05/mptcp_exp...
MPTCP's design was fundamentally incompatible with QUIC's UDP transport and session model.
From what I understand, neither MPTCP nor MPQUIC are designed to share multipath routing via a gateway (e.g. a bunch of devices sharing a multipath aggregation of 3 4G modems). It's all designed for a single device with multiple interfaces.
https://estore.peplink.com/categories/speedfusion-cloud/spee...
We use to use Peplink at a company I worked at
It’s surprisingly good.
I don’t think I have any special anecdotes except that I’m getting 85Mpbs up and a 35ms ping to speedtest.net.
Initially the connection wasn’t reliable, but it apparently needs to be set up for a few hours before it “locks in”.
This is on the edge of nowhere, but still cool. The only other option I had here was dialup or if I wanted to pay outrageous rates, I could get a .5-1Mbps T1 only because there is a ranger station not too far off that put one in and I can access some of the trunk hardware a few miles away.
- The setup is stupid easy. Place dish outside, plug in. That's about all.
- It's weird having a dish that points straight up, used to them all pointing south (hemisphere dependent of course)
- The router box (thing?) is a bad design. Stupid little triangle that tips over all the time. Seems cheap.
- Shipped with no wifi encryption. I thought that was weird. I bet this changes to for real release.
- Got a few "outside your home area" messages without moving. Errors.
- It is good for streaming or websites, but currently I have it on the ground and it seems like it loses connection fairly often. I'll put in up on a poll now that I know...
- The dish is heated. GOOD. I was worried about this with the amount of snow it's going to get.
I don't really geek out over these things. It's good. It's exactly what it should be I think. We'll see how it lasts the winter. I expect the reception to get better when it's mounted, but it'll be interesting to see how white-out snow effects the performance.
https://sebsebmc.github.io/starlink-coverage/index.html
(Edit: And evidently some regular users are already beta-testing it too.)
Amazing what passes for a minimalist lifestyle these days!
And soon 7 (!!!) internet links! Anyway, I just find it interesting how individuals (including myself) describe themselves and rationalize their descriptions.
It's normal to a point :)
Your priorities lies elsewhere, perhaps not in bowls, forks and knives, but certainly with internet connectivity :)
It sounds like OP has found that six different ways to connect to the Internet is the minimum that is viable for him, and that meeting that is important enough to justify spending a large portion of his limited budget (both financially and in terms of space in/on his vehicle).
If one fork does the job, one fork is enough. If five way to connect aren’t enough, then maybe six will be enough :)
A minimalist online meeting is an IRC connection over 3G.
They're just much more involved to get in a Van.
It was short lived magic as DSL came along. lol
I had no idea bonding was still a thing.
Also tested ADSL as a bond/fallback for a leased line, around then. Ah, the days...
> Yes, this setup is overkill but having functioning internet is especially important when going camping with young kids.
No, it's not. 100% no.
From the text however, its not clear how often the kids are around. Maybe they could benefit from some downtime, maybe not. Maybe it is their link to school and would therefore be even... crucial. In absence of info people tend to speculate. ;-)
Also, I have a stash of important movies that I keep for posterity. Even at a normal home with reliable internet. I don't want to pay for rental every single time, or have them go off streaming at random as they so often do. You'd be surprised what can fit on an SD card or micro usb flash these days.
Deliberately so, this information is private and personal. Especially right now due to lockdowns in Australia. I have publically shared this recently because enough is enough:
https://twitter.com/GeoffreyHuntley/status/14335483401221857...
For what it's worth it's interesting (but possibly explained by the ridiculous cost of anything not SpaceX) OP doesn't have a satellite internet uplink too.
https://ghuntley.com/how-long/
thanks for the spam
> If you are going to work remotely anywhere in Australia from a van, you need damn good internet.
In that context it may not be an overkill. If you're 2-3 days away from nearest wifi enabled spot, then having a solid backup strategy might be a good idea...
There's simply nothing better available, in 6 years of intervening time.
This world would be so so much more awesome if there were some really really good USB wifi adapters. This is yet another further leap, but I also think of all the guides to turning the RPi into an access point, via usb adapters: it starts out working fairly well, but the hardware either has hard maximum client limits or just fails to work well once you get to half a dozen devices. Even if you have something like PCIe or M.2 expandability, options are still extremely limited, & availability is incredibly poor.
AWUS036ACH is one of Alfa's most up to date USB cards. It is, however based off an RTL8812AU chipset[1], which is 8 years old now[2]. I'd love at least to have a good client chipset & equipment that supports WiFi6 that has good RF performance. I still think the market is overly structured & segmented. No one makes AP-class chips for USB, but to be honest I really think they could & should & that it'd be exceptionally useful & good. But no one dares to threaten the AP market like that.
Thankfully in wifi, RF engineering & antennas count for a huge amount. You still have great gear. But it does hurt me to see what seems like rampant tech-stagnation & neglect, of such a vital, interesting, basic part of modern communication technology: our wireless systems.
[1] https://wikidevi.wi-cat.ru/ALFA_Network_AWUS036ACH
[2] https://github.com/lwfinger/rtl8812au#purpose
Basically consisted of a gl.inet router that I plugged in a old Moto X4. This allowed me to get wifi from whichever campsite we stayed at (always garbage) and then I would pay for 100GB from Cricket and T-Mobile. I would swap sims when I would run out of data about half way through the month.
Then I discovered I could get unlimited data on Cricket with a regular phone plan if I changed the TTL on the router. Ended up being the most stable option for the last few months.
IIRC, on the plan I was using on Cricket, I would get unlimited data + 15GB of tether. This just byapssed that limit.
I think they can almost figure it out because if I tried to tether without the TTL trick, data wouldn't work. So I'm guessing it still counted the data I was using against some kind of cap?
I used to use them a lot when I was traveling with my laptop a few years ago and it always worked great.
But the good news is that I got it from the Play Store back then and you didn’t even need to jailbreak.
If you set it to 65 on your host device or router, it will be decremented to 64 on the phone - and is now hard to distinguish from real traffic from the phone.
You can tether with it and the only trade off is that you get deprioritized first in crowded environments.
There’s also no store, but I see that as an improvement.
Deprioritization can be hard in some areas. Verizon has the most customers and the least spectrum at the moment.
IIRC a long time ago all of Visible's data speed was capped at 5mbps and that was the 'tradeoff' of the low cost service in addition to deprioritization. They eventually dropped that cap, but it's possible it still exists for tethering.
I will plug Starlink for those that can get it though - not mobile yet, but we went from a crappy calnet connection to a stable 50+ sometimes 150+ mbps connection. When they get it working across cells (which Musk tweeted interest about) it'd be a pretty sweet option for an RV. If you're out in the boonies though and looking for stable home internet, it can't be beat (as long as you have a clear view of the sky).
https://excalidraw.com/#json=4810888066367488,n9qcxzAN-ynr7k...
I’ve followed her blog for years at this point mostly because I enjoy her diagrams and pretty much always learn something from the way she presents information.
I wonder why the author didn't consider a Peplink router which can handle redundant cellular/wifi connections out of the box. That's how I recreate a setup akin to this while I'm on the road in a single device.
Not to say this isn't awesome as is and I'd love an excuse to install a full Ubiquiti stack in my own camper!
I'm particularly interested in this topic because I've been thinking about power a lot after the Texas blackouts (considering building a bike generator so I can have power in case of a disaster like that one).
How does channel bonding work across multiple ISPs/NICs?
Edit: This seem to be it: https://speedify.com/blog/featured/speedify-protocol-mptcp-d...
"The way Speedify used to work is it opened one TCP socket over each Internet connection it used - e.g. one TCP socket over Wi-Fi / cellular / wired Ethernet / etc. With the new protocol, Speedify opens up multiple sockets over each connection (for example 8) - 8 over Wi-Fi / cellular / wired Ethernet / etc."
That seems a bit wasteful towards the server destinations.
The stronger the blur intensity, the amount of possible starting states to get to that blur increases exponentially. If you blur enough, every pixel is the same color, which obviously has destroyed all information.
But if I know that the blurred content is a social security code written in 12pt Times New Roman - I can perform a blur operation on a million SSNs, and see which one matches most closely to the mess I have on the screen. It's easier with bar codes.
Not true. The ASCII art people have sorted characters by their “shade”; if you know the foreground and background colour of a letter, and you know the single colour of pixel it blurs to, you can still work out what the letter was.
Blurring discards some information and obfuscates other information, but really, given how easy it is to reverse-engineer such things, we should measure the number of bits of information remaining.
If information in image + information about image > bits of information in sensitive data, then in theory you can recover the data.
So simply deleting the data from the image (e.g. blacking it out, removing reflections) is preferable. Saves you a lot of effort!
Totally agreed, and I like what you said about information remaining - you start to get into fun information theory stuff.
For example, let's say you're given one of the 'shades' you mentioned. I give you #a9a9a9. There's 32 bits of information there. However, if it's a blurred letter of black text on white background, it's always going to be grayscale. There's only 256 possible grayscale values - only 8 bits. Luckily, since there's only 26 lowercase english characters, we can easily fit that into our 255 values. Information is not destroyed! This is the shade->char map you were talking about.
But now what if we blur enough to turn two letters into one shade. That one shade will still be between 0 and 255. But now there's 26*26=676 possibilities that could've created that shade. More inputs than we have outputs for. There's no way to fit more than 256 possible inputs into 8 bits of information. Shade '98' could be 3 or 4 different inputs. However, we can be very clever...
We know that the letter 'u' comes after 'q' almost always. We know that 'x' never appears next to 'j'. We know lots of things, and can supplement the destroyed information, with outside information. We can actually change the whole context for this conversation now. We're not talking about blurring. We're talking about fitting english text into less bits than ought to be possible. We're talking about compression. This is exactly what compression algorithms for english text will do. There's a lot of redundant information in plaintext, just like there's a lot of redundant information in our images of text. An 8x8 pixel character glyph can easily reduce the information to a single pixel. However, there are limits. You can compress english text by 10x, if it's simple enough. You can't compress War and Peace into 5 bytes. You can blur text a few pixels and not destroy information. But you can't blur a paragraph of 12px font by 500px and get your original information.
It’s difficult or impossible to derive the original information from the output, but if you have a small enough keyspace you can generate all of the blurred (“hashed”) versions and compare the results.
https://i.imgur.com/WeKOofK.jpg
I never really stopped to ask this of my phone either, I guess.
Hold on: heating up part of the human body that are not capable of dissipating heat efficiently can increase cancer risk as well.
Yup, exactly. Was wondering if there was any data in the evidence medicine based community to be concerned about. Sounds like there isn't.
https://scholar.google.com/scholar?hl=en&q=microwave+exposur...
This is not to say that low power microwave exposure is worse than exposure to pollution, unhealthy food, risks of car accident.
But it would be incorrect to say that microwave transmissions are free from risks.
The mere presence of studies on a subject does not indicate positive evidence of material risk to health.
So far, nearly every redundant link I've found seems to lead to bugs and glitches all round when failover actually happens.
(I have a similar setup to the OP, but using an array of M.2 4G/5G modem cards connected to external antennas.)
Any guidance you could give on the most bang for your buck way to get internet?
Just a hotspot?
Did not know about the Nighthawks!
Those are frickin sweet!
Thanks.
I use a Pi running Easytether connected to a 5G phone, with a travel wifi router connected to the Pi over ethernet. Runs my off-grid internet + security cameras.
What I think most of these solutions do is use MPTCP or a custom homebrew protocol to a server that then holds the actual TCP connection you want. So a slightly more complicated VPN. There is an open source project for this:
https://www.openmptcprouter.com/
There’s more than one choice in the space.
Peplink also has their own but it’s geared towards businesses and is pricey.
Based on other comments here, I think that's exactly what Speedify is.
Also you can simply use multiple uplinks without VPS and it provides better latency. The drawback is that you can't failover/balance protocols that use only one connection. Mosh and browsing work well tho.
From https://speedify.com/features/
> What kinds of servers do you have? > We use a combination of dedicated and virtual servers, depending on location and scalability needs. All of our servers have at least 1 Gbps network access and are optimized for bandwidth intensive usage such as live streaming, video conferencing or online gaming.
Let's assume the best case, where Speedify implements this as swarms of redundant servers, so that even if some servers or datacenters are down it may continue to work.
I could not find in the article any mention of disabling Speedify and falling back to using a single provider, or a local bonding solution like described on https://wiki.debian.org/Bonding .
Isn't speedify still a "single point of failure", in the sense that any incident with Speedify (e.g. accounting glitch of any kind) may cause the whole contraption to fall like a house of cards?
1. https://www.abc.net.au/news/rural/2020-05-01/rural-residents...
2. https://www.gizmodo.com.au/2020/05/telstra-sms-over-wifi-lau...
3. https://www.itnews.com.au/news/experts-call-for-rethink-of-c...
Your comment makes me realize we may need more options on what to do when the connection fails. Security-types had us add a "kill switch" to make sure traffic couldn't get out when the encrypted tunnel fails. But if you don't care about that, another option to guarantee it would go right out, would make sense too.
I'd be interested in paying for the service but not willing to run closed source software locally and then send all my traffic to it.
Just wanted to drop in this thread and let you know I think you have an awesome product and I have been happily using speedify for almost a year now. I hope you keep up the great work and are able to keep growing!
Most of the time you can simply route traffic through multiple uplinks without bonding.
If you are on a Zoom call, tunneling through Speedify in redundant mode will give you the minimum latency of n connections, reduce jitter (latency deviation), and eliminate packet loss.
Whatever solution I researched it boiled down to trading bandwith increase for latency increase.
Speedify seems intriguing yet seems extremely country dependant.
Say you are in Sweden and want to access Australian site when Speedify server is in UK your latency is surely to go up.
Since Speedify is essentially a VPN with multiple incoming connections, isn't it extremely important on what country Speedify servers are?
PS Another up/downside your public IP is probably Speedify IP (just like a regular VPN)
On the software-only router side of the house, Speedify at least let’s you tie your multiple links together and get you a more reliable connection. But true aggregation (like LACP) requires that all links be of exactly the same speed — e.g., you do link aggregation on multiple Ethernet ports that are connected to the same switch, and while no one connection will have bandwidth greater than a single Ethernet port, you will have be able to have multiple connections that can all achieve that same level of speed. Their style of bonding isn’t the same thing.
I’ve also looked at OpenMPTCPRouter. On paper, it looks like a better solution to me than Speedify, but in practice it’s not designed to handle multiple different upstream connections with varying levels of performance. I’ve chatted with the author, who lives in France and his ISP connection is 1.25 gigabits per second, so this hasn’t been a priority. He did say that this was going to be something he was going to start working on soon, however.
So, I might be forced to go with something like a six port Protectli pfSense or OPNSense router plus Speedify, at least until I can get something better.
So, you’re not going to tie multiple upstream sub-1gbps providers together into one giant network connection, because Speedify will just stop bringing up more channels when it thinks you have enough speed.
At least, that was the Behaviour I saw when I was doing my tests. Maybe I had the wrong type of account, so that they were effectively throttling me down?
I was able to get some higher speeds with Speedify than I could get with my naked ISP or with NordVPN, and the graphs showed that the cellular 4G connection was actually a significant bonus on top of my DSL.
So, now I guess I’m going to have to build my own Speedify router combined with OPNSense or pfSense, and get those other upstream network connections also set up.
https://www.nokia.com/networks/products/fastmile-5g-gateways...
In fact, their website indicates that they're customizing the device with differing specs for different customers with options like an LCD touchscreen and differing antenna designs.