I'm curious, how they actually be able to reduce the transfer costs from AWS? Do they know some smart way to avoid paying to Amazon for the transfer and still be able to get the data out?
R2 is still under development, we will see how it will work, but I can tell you how Tebi.io global replication works:
- Writes and reads always goes to the closest region, this keeps network latency low. You can define different metadata write concern levels for each bucket, this way you can define how fast or how consistent your writes will be on a global scale. You can even make them asynchronous, meaning that once the data transfer is complete, metadata write and propagation is performed without you waiting for it to complete.
- If you write to one region and someone is trying to read that object from another region - it will be possible as soon as metadata is replicated to that region (usually it takes less than 500ms). If data is not yet replicated to that region, then it will be read from another region. If the data is partially replicated, then that part will be read from the local storage and the rest of the data from another region. Additionally, Tebi supports synchronous replication allowing almost instant data propagation across the world.
- Write latency depends on metadata replication write concern - it can be faster than AWS or slower, you can configure it yourself.
- You can define where you want to store your data and how many copies you need in each region.
Check tebi.io - it is a geo-distributed S3-compatible storage that does exactly that. You can configure global consistency level for each bucket individually.
> Our vision for R2 includes multi-region storage that automatically replicates objects to the locations they’re frequently requested from.
Seems like they are going to automatically replicate data to other regions. Something like tebi.io is doing for a long time already, it is a geo-distributed S3-compatible storage that is replicating data across the world to reduce read/write latency and increase speed.
If it is done right, this might increase download speeds by a lot especially for big infrequently accessed files.
This is true - most of the object storage providers are really bad at serving data fast, they more focused on a cold storage. Tebi.io works really fast for small files, plus it is a geo-distributed object storage meaning that data is physically stored in different locations around the world, greatly reducing latency.
Backblaze, DO Spaces simply were not designed for this in the first place.
You need to account IO load as well, more bandwidth means more requests to the disks and I assume they will use mechanical HDDs for the storage, SSDs are still too expensive to support $0.015/GB. Good mechanical drive can give you up to 200 IOPS and maybe 200MB/s, so they will have to copy the data a lot to be able to increase capacity on demand. Making it free doesn't seems to be sustainable at all.
I suggest to use constant quality encoding instead of a constant bitrate, this way encoder will automatically adapt bitrate for a particular scene. This is a much better approach, it will give you better quality and a smaller file at the same time.
For example, encoder might choose 1Mbps bitrate for a scene which is mostly static and 20Mbps for a part of the video with a lot of movement. Your 8Mbps constant bitrate will be an overkill for the first scene and too low for the second. Let encoder decide the optimal bitrate.
Actually many companies are trying/tried to compete with AWS especially in an egress price territory but because of a lack of exposure it is very hard to get a traction and get big enough.
And there is a lot of things that can be improved - like egress price, better support, get rid of regions completely, etc...
Like: They are pushing the web and technology in general by inventing a ton of different services. This is really nice.
Dislike: Traffic price is just crazy high, it is simply impossible to start anything that is bandwidth-heavy, they might give you a good discount after you spend your first $100k but still there are many alternatives to the most popular AWS services like EC2: Digital Ocean, Vultr, OVH; S3: DO Spaces, Tebi.io, Wasabi, etc.
Doesn't seems legit at all.