What if the content is just generally in "bad taste" and not overtly Neo-Nazi, will CF feel the need to play "Content Cop" or are they willing to abide by their role as utility?
Is loss of human life an acceptable cost for free speech?
Tech companies didn't choose to be the society police, yet here we are.
Name one person harmed by the site?
If you have ever seen a page of DailyStormer, you'd either laugh or go away.
No one has ever killed anyone after reading a few paragraphs on that site.
Putting in the argument of "Is loss of human life an acceptable cost for free speech" is not an honest way of describing censorship of content that is silly at its worst and dumb at its best.
We just witnessed something very significant, when Russia interfered. They had an express goal and they carried it out with technologies we know.
A medium that is absolutely free of censorship is not in your best interest, because it can and will be used to compromise you.
The solutions are difficult, but what is clear (especially to folks who have had to moderate even a small channel or forum) is that a policy of no-censorship never scales in the real world, the question is how you do it, where to draw the line, etc..
Yes, this decision has already been made. Freedom (of speech and more) is fought for and won through struggle and sacrifice. We have already lost much human life to defend these rights and continue to do so.
Bad actions will be dealt with the appropriate consequences, as they always have, but that does not mean we should start limiting rights. This is the foundation of the justice system and society itself.
https://www.theepochtimes.com/anti-terrorist-hacker-group-re...
Just because one has freedom of speech does not mean one is free of consequence. So actual murderers should be punished. People believe what they choose to believe, and if it's not the DailyStormer it's some other hate mongering source (that's not exclusive to the internet).
lots of people would say yes.
A recent example of what appears to be a moderate, democrat voting professor experiencing it:
http://www.dailymail.co.uk/news/article-4546704/Professor-ca...
There is a huge gap between “conservative bent” and those trying to actively incite violence in the name of Hitler. And that they’ve only removed ONE such site across all their hosted properties is hardly an indication that their CEO is randomly moody. A single data point is not a trend.
I know many YT firearms channels that have been basically blacklisted from monetization in recent months because YT classifies anything they do as "unsuitable for ads". It has forced a lot to double publish to https://www.full30.com/ to monetize their work.
Trust is easy to lose and hard to gain.
>are they willing to abide by their role as utility?
I can remember when the conservative position was to support freedom of association, and to oppose regulating private businesses as utilities.
I can remember that as recently as this year, in the debate over net neutrality.
The nazis are shit too, no doubt, but it’s hard not to blame the antifa for the ultimate outcome.
Also worth remembering: the nazis were using their legal right to free speech in a legal demonstration. The antifas were illegally and violently obstructing it.
Don’t be too quick to pick sides.
Surely, this piece of the pie is not as big as many had hoped for, but there’s definitely more than a 1000 companies that think their video content is worth a better platform than YouTube or Facebook, however they define “better”.
What’s the deal with ‘lossless compression’? I feel like they use the term very differently from how the industry uses it.
Multiple resolutions, adaptive bitrates, dynamic packaging for different versions of HLS/DASH/Smooth are pretty much a must-have for any video solution these days, free or paid.
It’s amusing to hear an argument that CDNs are hampering adoption of better video compression from a CDN company. However, while Akamai surely would love to bill for more bytes per minute of video, no one is asking them. They just distribute whatever and however many bytes origin server has for them (give or take some convenience features around it).
CDNs are scrambling to provide compelling features to increase stickiness, usually with limited success when it comes to video. Bandwidth-heavy customers do want to take advantage of rapidly commoditising technology and falling prices and are pushing for multi-CDN strategy.
Using correct terminology and customising stock Bitdash player would have helped at the start of such ambitious endeavour. Good luck to CloudFlare and congrats to Bitmovin.
Thanks, I contemplated whether I should disclose it, but given that I have currently have no stakes in the game, decided it's unnecessary.
Edit: And a related question: would Cloudflare Stream work without the JS player using a regular video-tag?
Depends on the browser. This is the primary reason why people actually use a JS player that wraps around video tag. The other reasons are to accommodate business logic (i. e. overlay on top of the video, pre-roll, thumbnail, etc.) and better UX.
Other CDNs charge for bandwidth so why would they offer their customers an easy to use 'optimize images' feature, it'll just lose them money. Cloudflare can offer it because making your images smaller saves them money.
This will be an interesting option for anyone who publishes videos but speed, integration complexity and pricing are going to be key.
The primary use case for CDN is latency reduction for serving small files (CSS, JS, images) and such traffic has different characteristic and requires different optimizations than high-throughput traffic. I worry how long large files are going to be cached by CDN PoPs. PoPs can't certainly cache everything forever and they can be more eager to remove large files and in such case CDN will be reducing performance by adding an additional HTTP level hop to the origin server (and likely a few IP level hops). My other worry is that PoPs can deploy DoS detection heuristic that are tuned for standard web pages and a bandwidth intensive page may trigger the rules and have the traffic throttled. Such throttling is something that I've occasionally observed while testing CloudFlare with large files, but I'm not sure what was the cause, CloudFlare monitoring tools did not report anything unusual.
Check out cachefly.com for instance.
One of the challenges with video delivery and where sites like YouTube can't work is if your business relies on providing videos to paid or logged-in members only.
Patreon creators get around this in a kludgy way by using unlisted YT videos (which can still be shared by rogue patrons) but that's not ideal.
But if you want to control access to a stream based on some custom business logic in your application, you're currently stuck serving your own video or using one of the existing combination of encoder/server/player that their post alludes to.
Pretty hefty work involved in getting it reliable, would be nice for a "drop in" solution.
Imagine just dropping a tag like this onto your page:
<display-ad width="300" height="250">
Meanwhile even Google has to crawl your Adsense pages to know what sort of content to serve. And they have to worry about things like websites serving different content to GoogleBot. Some interesting possibilities when you're proxying all of a website's traffic.
Even if Adsense let you render ads server-side by proxying user info to them, that's a lot more work for every website owner compared to what CloudFlare could enable since it already is the proxy.
Or imagine CloudFlare encoding live-auctioned ads directly into your videos on demand as part of the video stream itself.
As a cloudflare customer and an everyday internet I am thankful cloudflare hasn't gone in this direction.
So the statement that "the video streaming market is screwed up" is based on the fact that Cloudflare themselves only have 1K customers doing a meaningful level of streaming?
You might want to talk to Akami, Amazon, Edgecast and Fastly. I'm sure they will tell you the video streaming market is quite healthy.
There are plenty of places for video, but I am not so sure we need more of it. I'd say the tendency to put more content into video format is an overall negative trend.
I happen to work directly in this space and I want to know if you think my traffic numbers are meaningful.
I'm not in a position to explore a need for Cloudflare outside of hiding my origin IP addresses, but for those of you who are, are you actually using and benefiting from the wide array of offerings from Cloudflare? As it stands now, I see these announcements and make a mental note to expect a slough of marketing emails in my inbox for the next couple weeks.
Why would you want to losslessly compress audio or video? The resulting files would be huge--often too huge to steam. Lossy compression is what you want.
I must be misunderstanding the article?
That's a slightly odd way to phrase it, and I'm amazed that it's not a basic part of any streaming service these days, but makes more sense to me than the other suggestions.
For instance, the user might upload an H.264 video as an MP4 file. Streaming services generally won't serve the MP4 file itself, but will create an HTTP Live Streaming (HLS) version where the video is split into multiple .ts chunks along with a .m3u8 playlist.
However, such services (including YouTube) will generally re-encode the video when doing so, and the end result will have poorer quality than the original uploaded file.
This is actually unnecessary because .ts files support the H.264 codec, so the video does not need to be re-encoded. Instead the streams can be copied to retain their original quality.
If additional compression is necessary, I suppose it would be possible to apply HTTP compression using algorithms such as gzip or deflate.
Of course, this is purely speculation. It would be nice if Cloudflare clarified what they meant. But it would certainly be great if they used stream copying where possible instead of re-encoding.
1920x1080 at 3 bytes per pixel and 30 frames per second is 11GB for just one minute of video, and 186MB per second. That's far too big and expensive to stream, and outside of somewhat special circumstances I've never heard of anyone storing video in such a huge format.
- your product is explicitly lossless as a feature (like Tidal for streaming lossless music)
- you have to provide the files to others for reasons other than watching or listening (like editing), where integrity matters, and the use case/context requires a stream instead of just a file download
An easy but somewhat crude comparison would be a text file that reads:
“Space,” it says, “is big. Really big. You just won’t believe how vastly, hugely, mindbogglingly big it is. I mean, you may think it’s a long way down the road to the chemist’s, but that’s just peanuts to space.”
Lossless would be putting that content into a ZIP file (eg winzip / 7zip / etc).
Where as lossy would be rewriting the passage to something like:
“Space is really big. You just won’t believe how vastly, hugely, mindbogglingly big it is. You may think it’s a long way down the road to the chemist’s, but that’s just peanuts to space.”
Generally lossy compression will yield better compression because you can vastly reduce the information you need to compress and then apply lossless techniques as well.
Of course lossless compression is typically a lot less compact than lossy alternatives (the article mentions a 1/2 compression ration, which is not much compared to lossy video codecs) but it means you're not degrading your video at all, you just have additional processing.
Note that some algorithms these days claim to be "visually lossless" which is mainly a marketing gimmick which means that it's lossy but you wouldn't be able to tell the difference with the naked eye in normal conditions (TICO claims that for instance).
In ffmpeg, -qp 0 will get you lossless x264. I use it all the time (via OBS) to losslessly record my screen.
For instance try zip'ing a WAV file and compare it to the same file encoded in FLAC. Typically FLAC will achieve better compression while still being lossless.
There are lossless video codecs, like VC-2 for instance. It's not very common in the consumer world however because people generally don't want to stream or store high definition video with a single digit compression factor (that single digit being generally around 2).
I know a few online businesses that don't use YouTube because they operate in a grey area in terms of fair use on videos (videogame footage) and they don't want to build their video business on YT because of their "trust the claimant over the content producer" and 3-strikes approach to DMCA claims.
Obviously if customers are paying Cloudflare to host their videos, I would assume the dispute process would be less one-sided, but have they said anything about this yet?
- will Cloudfare require a legal DMCA takedown notice in order to take down content, or will they allow 'trusted' partners such as large companies to simply assert a violation?
- will Cloudfare employ any sort of automated copyright screening algorithms, like ContentID, that automatically prevent sharing of some content?
- when served with a DMCA takedown, will Cloudfare immediately take down the content (presumption of guilt) or will they allow the content-provider time to respond, either to contest the takedown or take responsibility?
- will Cloudfare push back against or provide consequences for filing fradulent DMCA takedowns?
>So, will Cloudflare have the same kind of DMCA and/or ContentID controls that YouTube has? How will they police the content they stream for their clients/partners?
On top of that, I suspect there's no "3 strikes" rule, either. A cache is a cache. Entries in the cache expire for a lot of reasons, including legal ones. Lots of cache expiration has no impact on ongoing caching of new data.
Imagine a world where all ISPs didn't let you spoof IP address. And, controversially, one where ISP customers had to pay for their SmartToaster's outbound DoS.
Edit: Seriously, -1's and an accusation of supporting pedophilia? I've certainly hit on a raw topic. And frankly, I'm not wrong regarding Captcha hell. We've had dozens of articles and threads about that very topic.
Cloudflare still has a long way to go in terms of being nice to humans.
For a short while, it was significantly better. It probably lasted for 3-5 weeks, and then things fairly quickly progressed back to their usual, namely being captcha hell.
(Request first page, do captcha, click on next link, do another captcha, go back to first page, do captcha... wash rinse repeat.)
A system like this will make it easier to create Youtube competitors, so when youtube screws over a part of their producers / consumers they have somewhere viable to turn to.
How they deliver the video is pretty innovative:
> How we accomplish this is by dynamically assigning data centers IP ranges based on which are currently being underutilized. Then videos are loaded alongside a manifest file, pointing the player to where they can locate the different chunks of the video for each bitrate. (As we described earlier, this is how adaptive streaming works: the player uses this manifest file to quickly locate different bitrate formats and switch between them as network latency changes). That manifest file is dynamically generated to point to the closest data center using Cloudflare’s default Anycast IP for the first chunk, then to the underutilized data centers for subsequent chunks. The final result is that the first segment is fast, with low cost subsequent segments. This allows us to offer a video solution that is at the best price, because let's face it, enabling video today is really expensive and unpredictable and it doesn't have to be.
Dynamic Manifest generation has been done for years, first by Unified Streaming, now by most companies.
YouTube is, essentially, free. Even if I don't want to use them as a host, I can upload a video and a few minutes later download it in DASH and a variety of older formats - again, for free.
If I do use them, I might get a few quid of advertising revenue.
CloudFlare video lets me keep people on my site (damn those distracting cat videos) and possibly gives me a branded player. That's nice - but is the cost of use (per minute transcoding) going to be worth it?
eastdakota: "We plan to support live streaming. We're gathering data on how people would prefer to do streaming. As soon as we figure out the best way to support the greatest number of use cases we'll add it to live to Stream."
Definitely really interested to know what they are thinking of doing. Potentially a customer if they do it right and much cheaper than the competition!
Linked from this page: https://www.cloudflare.com/performance/
https://techcrunch.com/2013/04/08/yc-backed-tapin-tv-evolves...
Sadly, it didn't work out for a number of reasons, largely unrelated to the actual idea. I'm glad someone else is finally working on it.
> know what a megabit per second or H.264 are.
Yes, yes he should