That said I don't hear the best about the company (long working hours, not the highest pay) and have declined a job offer there myself.
As someone who lives in a low cost country: rubbish. There's a reason we're a lower cost.
1. We have a lower standard of education
2. We have a higher cost of technology (relative to average income)
3. We have a lower need for the luxury market where most technology resides
You can also look at the proportion of field leaders. Are more from the developed nations or the developing ones?
The developed nations had a headstart on technology, do you think the developing ones have overcome that, despite most of us going backwards in terms of access to education and the wealth gap?
Don't take it personally, I'm not trying to tell you that you're a bad developer. What I'm saying is we have to work harder to uplift our country's fields and not fall into the trap of "well I'm the biggest fish in this tiny pond so therefore I'm an equally big fish in the ocean". It does not work that way.
Lower cost of living does not imply developing nation.
Czhech Republic or Poland or Taiwan are developed nations, all with the cost of living a fraction of Bay Area.
I see Ubiquiti dev center apparently moved to Latvia. You can argue it's a depressed region of the EU but it is not a developing country by any metric.
A lot of the engineering talent you find there is extremely limited in quantity, but it is untapped.
Overall I think it’s a step in the right direction as compared to other countries we’ve outsourced work to, but let’s not pretend Poland is an exemplary ray of developed industry and society.
Sure, if you take 300M market like US and pool all the best talent to west-coast, there is a lot of engineers. But the market is saturated, and it's nigh-impossible to hire a team of the magnitude you can get in Eastern Europe. Canada? Oh please...
Having worked (as a French) with dev teams over there, I can assure you they have a whole bunch of A+ devs and the market is very competitive.
You don't have problems with access to clean water in Poland. You ain't going to die from hunger in Czech Republic. There are no issues getting education in Latvia.
And mythical Anglo engineering talent, please. They are approximately the same anywhere you mentioned.
But I wouldn't think of the Baltics as such a place?
Instead, just hanging over the code and infra to new people, is risky in itself and can easily make security problems happen?
I'd like to also add that from my own perspective Ubiquiti gear and software is/was _exceptionally_ good. So converting to "just another network gear company" level of quality is going to have a much greater impact on my perception of the company compared to if this news of off-shoring came from say cisco.
In the US, good engineers don't have to tolerate abusive companies in exchange for good wages. Not all countries are so fortunate.
Almost forgot, but just last month my wife started complaining about the WiFi, turned out the latest firmware that was pushed to my UAP's is horrible. You have to turn off auto update and check reddit before upgrading firmware, what a joke!
> I hope it's not "the company is declining because the company moved talent to countries where developers are cheaper"
We probably agree the cost of developers varies considerably in different locations.
Companies operating in markets where local developers are expensive (I work near London for example) sometimes decide to outsource development to locations where it is much cheaper (e.g. India).
What I have seen (admittedly just anecdotally, not carefully studied and subject to bias) is a correlation between companies that don't value high quality software development and companies that are happy to dump a substantial part of their development efforts on cheap developers overseas.
Such companies don't care about their existing team - developers who are expensive but have built the software and understand it, and are to some extent understanding the end customer (often filling gaps left by product managers in smaller organisations). They just see their development team as a huge cost that needs to be reduced. Or they have trouble hiring locally.
When the daily rate of developers overseas is substantially lower, the company also doesn't care about those developers. They will want to outsource the least creative and satisfying work to them. They want to just fire and forget: "here's the spec, go and build it". Lower productivity is less of an issue if the "resources" are cheap. I'm over simplifying and I'm sure some companies try to work more in equal partnership but you get the picture.
Also, the move to outsourcing is often done suddenly; it's more like wielding an axe than an organic growth or shift in development model. Companies don't tend to invest in overseas developers as individuals.
In my view it's not great for either set of developers, or for the customers of that company. Perhaps when done right it can be more beneficial and I hope things will improve overall.
You cut my quote short of the important part which was "and employees complained less about constant crunch mode". It is easier to "trap" engineers in countries where average salaries are lower by offering them 20% over market rate and then threatening to fire them if they don't work constant crunch hours. We were promised very large bonuses that never arrived.
> That said I don't hear the best about the company (long working hours, not the highest pay)
It is a sad situation. The pay was very good and the working hours were reasonable when I started, but that changed for the worse. That is why all of my peers left the company.
My company started doing the same and erosion of knowledge is really bad.
You are saying the same as the parent. Expecting “short” hours is seen as “feeling entitled”... There are countries where people are not willing to work in those conditions.
Bottom-of-the-barrel developers bring even greater savings, and paying peanuts has always been a great way to get monkeys.
Are there any other "prosumer"-type devices on the market that could replace a Dream Machine? If Unifi is going downhill it doesn't seem like I'll be going with them for a replacement.
Opnsense forums have lots of recommendation for hardware, which is the path I went recently. I went with https://protectli.com/, which are just some rebranded hardware sold on Alibaba, but they provide support ontop of the hardware.
I actually reversed this choice and am back to using the Unifi Controller again - pfSense is superb in production or more-networking-enthusiast style environments, not so nice for "average" home. I used a 5-ethernet port fan-less Intel Atom box almost identical to the one you linked for my homemade pfSense router while it was running, for that purpose it was pretty good.
But at the same time, I run Google WiFi points as I don't want to deal with them. :)
The benefit of UniFi is that you can centrally control a bunch of switches. It's definitely overkill and overpriced if you just want an all-in-one.
I keep hoping that one of the OpenWRT snapshots will fix it, but this is something I've been fighting with for years on multiple pieces of hardware.
Separate 5G and 2.4G SSIDs are no issue in my experience-- it's multiple SSIDs on one phy where Atheros goes wrong.
I really want WiFi6 gear... but given that there's no real OpenWRT support yet, and how long it's taken 802.11ac to mature in OpenWRT (it arguably hasn't yet), it's kind of discouraging.
maybe you have a known-bad hardware in regards to driver support, but that is absolutely not to be expected.
I had been considering trying out some of their gear, but perhaps I might not bother now!
I replaced my ISP-provided all-in-one box (Orange in France) with an EdgeRouter 4 and it is many, many times more stable. The crap you get from the ISP does not compare.
The management is horrible and how they designed it is horrible as well (from the OS perspective) but once it works, it works.
Data leaks happen! It shouldn't but that's just how the world is. UI has been honest about it, and informed every customer as a precaution. (I assume they're still investigating).
I can't be sure, but since UniFi Video went offline at the same time the breach was announced, a week earlier than it was scheduled to, that might have been the entry point.
In any case, the UDM (despite all the negative talk) is a fine machine, and does what it promises to do. If you want similar performance you're either looking at building something yourself, or paying twice of what you paid for a firewall appliance. The Netgate SG-3100 has less performance at twice the cost.
You need a UI account to set it up, but that doesn't mean you have to allow managing it from the cloud. Disable the cloud controller access and any access to your firewall configuration will have to happen from your local network. I'm unsure if you can disable the UI account, but i have a spare UDM sitting around so i will test it.
What disappointed me is that some aspects are really unfinished, and it looks like there's no intention of it to be fixed.
For example we bought their pro firewall (which has been out for years), it's got 2 WAN ports for automatic fail over. To use the 2nd WAN port I had to switch over the UI back to legacy mode. Ok weird but I guess the new UI is still sort of new. But then it turned out that to configure automatic fail over in the most common way, I needed to ssh in and edit configuration files manually.
It didn't turn out to be very hard, but it was just jarring. One of their flagship products, and of the 4 ports it has, 1 port is not supported in their main UI and it's most common use is not possible even in their legacy UI.
Unifi Protect has similar incompleteness issues.
I don't think there's a company that does it better than Ubiquity right now, just disappointed that it stops there.
I've been running Ubiquiti gear for years, from a single 2.4GHz UAP with the Edgerouter products, to my current setup with UDM Pro, 10 GB backbone and multiple NanoHD access points, and to use an Apple quote, "It just works". I don't have a complicated setup, just some basic VLANs, firewall rules, radius assigned VLANs via MAC, and IDS/IPS, so maybe that's why i'm not having any issues with it.
I have the technical skill to set it up from scratch if i wanted a second day job, but i don't anymore. I've run on homebuilt devices, on a Soekris net4801, on an Alix APU1D4, on m0n0wall and PfSense in various configurations, latest on a Netgate SG-3100, and while the SG-3100 comes very close to being a network appliance, it still managed to crash to a point where i was flashing it and setting it up over a USB cable, and while Netgate support was very helpfull, that's hardly something you'd ask the average consumer to do.
On the access point side of things the only real contender would be Meraki, but those are 2-3 times the cost of UniFi gear. You could of course also get a bunch of Zyxel/Netgear/whatever consumer devices and put them in bridge mode, and lose all central management.
I would generally expect the UI to be for enthusiasts, with the more advanced functionality hidden in the CLI (kind of like Windows). WAN Failover probably isn't super popular among enthusiasts
I used to have several Ubiquit USG devices as well as their EdgeRouter.
I moved to pfSense as it's open-source, more stable, and gives you much better control/configurability on your hardware. There's a great ecosystem of packages on pfSense, that you can install via the web UI - making it a really feature-packed for a homelab.
However, recently I've been moving to VyOS to pfSense, which is basically a stripped-down Linux distro, with a heavily tuned FRR routing stack built on top of it.
VyOS is an open-source fork of Vyatta, which was previously owned/released by Brocade networks.
It operates with a CLI, like many enterprise/commercial routing products. It takes a bit of getting used to, but it's really great to use in practice, and makes it easy to diff configurations, or rollback changes, or copy the same configuration across multiple devices.
And of course, it implements with config-management software like SaltStack/Ansible (via Napalm), which is something that pfSense. If you have multiple pfSense devices, you basically need to point/click via the web UI on each one.
For APs - Ruckus is great, as is HPE/Aruba (they have a new low-cost line that's targeting the Prosumer market) - they have both been leaders in the wifi field for ages, and have things like AP handover, RF tuning/optimisation, adaptive antennas etc down pat.
The wiki is good and the community is really friendly. If you have networking experience or want to something to tinker with it’s a nice deal. If you want something you can set and forget I’d look elsewhere though as the UI is not friendly at all.
Their WiFi APs are behind the curve (no mu-mimo even afaik), but you can just hook up some other wAP if you need the newer protocol features.
What I do is keep a Mikrotik router that does all the heavy stuff and hang wAPs off of it as needed. I especially love capsman for wAP management. They do have all-in-ones of course, just not my cup of tea.
The hardware is indeed a bit lagging, I'm not arguing there, but it's not always (only) the hardware that makes prosumer.
hardware stats and software features are severely behind the curve but it mostly just works and you dont really need those features anyway.
certainly not prosumer
It is on the expensive side, but the hardware is beefy and you get vendor's support for OpenWRT out of the box.
Did you try? i did.
The controller UI shows you a hole in the left part of the diagram and explicitly tells you "no routing control without USG"
I have several Unifi switches and a controller (running on an rpi) on my network but I use my own router. I can setup VLAN access ports and trunks all day on the switches no problem, but I can't control the layer 3 routing between those VLANs with the controller, which is what you're talking about. By setting up a gateway/network on each VLAN from my router I can control routing. It's just not as slick as having a USG where it's all controlled via the controller UI.
It is not all nicely integrated together if you use a separate router (obviously), but it's not like it makes it impossible. It's not even difficult... at least not any more than it would be in any other setup.
I’m a big fan of the ecosystem and I’ve recommended it to many people but I’m constantly astonished by the slow pace of hardware updates.
I am aware of the UDM Pro and USG Pro but those things are expensive 1U monsters. Maybe fine for SMB use but this is for home use and I’m very space constrained.
If Ubiquiti made a small footprint security gateway with some modern hardware (the USG3P is some 8 years old at this point!) I’d buy it in an instant.
Not sure where to go next, but it probably won’t be Ubiquiti.
Robert Pera (the CEO) got his start in the industry in San Jose.
1) You start a .app that sits for a few seconds then requires you to launch the browser by clicking a button. While using the browser, you can't close the extra window for the controller.
2) On the browser, you go to a localhost website that has an invalid TLS certificate (you can a "Not Secure" warning) and have to click through to the unsafe website (and it's still like that in my current Unifi version).
3) The login page doesn't let you use the Chrome password manager, so you have to type it all in each time to access a local program.
4) In the web UI, the icons are not intuitive, and some combination of circles and rounded rectangles.
5) The new UI makes it seem like you can configure things that can't actually be configured outside your router.
6) Speaking of your router, Ubiquity's own EdgeRouter routers aren't supported in the Controller UI. They require a completely different interface.
In case anyone thinks the problem with the certificate is something to do with my own setup, it's not. It's a universal problem [https://help.ui.com/hc/en-us/articles/212500127-UniFi-SSL-Ce...]
I haven't had the password manager issue you describe. KeepassXC in Chrome and Firefox both fill out my credentials successfully on the login page. I totally agree about the UX of the web application though. It feels like over time, options have become more and more hidden and the icons more cryptic.
- Without a valid SSL certificate, there's no way to tell whether you're actually visiting your UniFi controller or a honeypot. Ubiquiti isn't the risk here.
- UniFi features that depend on WebSocket and WebRTC are unavailable when using self-signed certificates. This includes live stats updating, device terminal, airView, etc. (Those features can be used in the cloud UI... if your Internet connection happens to be working fine.)
- Valid SSL certificates would be easy to auto-provision these days with LetsEncrypt. There are some minor challenges around port forwarding / relay, but that isn't rocket science. If Plex can figure it out, Ubiquiti can figure it out :)
What to say - maybe before assuming that someone "don't really know how any of this works" you may, just a second, think that the person your comment is directed to has written a security reverse proxy and presented on that on one of the largest security conferences.
Or not, maybe that I really do not know how terminating traffic on MY reverse proxy and sending it upstream to MY ubnt controller works. Who knows.
I too have a working reverse proxy setup or few. I certainly don't expect something using a "localhost site" to come with valid certificates. Unless they somehow get a valid cert for https://localhost
Edit: apologies for the assumption, I didn't realise that you weren't the guy I originally replied to. I'm new around here.
Here's a review: https://seabits.com/teltonika-rutx11-lte-router/
I once worked for a company, there were some "grievances" between the programmers and CEO (nothing major) but enough to elicited a "meeting" between all the devs and the CEO to "smooth" things over and build a better path forward, we will all in high spirits for the meeting and optimistic.
The very FIRST opening line from the CEO in the meeting was:
"How extremely lucky we(programmers) are to be working there..."
It all kinda just went downhill from there... 6 Months most of the programmers quit.
It's not traditional networking gear, sure, but I can certainly see the play they're making, so I wouldn't call this a scatter-shot approach.
But I think this is an example of them having an ideas they only end up abandoning. They are now selling an access control solution, but would you trust them to be standing behind this in 5 years? Enough to deploy it to a customer building? Not me. Attractive for the hobbyist perhaps.
openwrt is based on Linux
I hope things are better now.
He's not wrong
Which is pretty hard to disagree with.
The saddest part is that we had many good engineers who could have continued to do amazing things with the UniFi momentum. So much time was wasted on dead end products like FrontRow. Most everyone I know left for jobs where we were treated better and paid more.