Good morning.
Let’s suppose that a new customer signs up, starts using it and the number of write and read operations goes up until it reaches a quite high level, let’s say 8 billion per month.
What would happen? Would he be kicked out? Would he be charged for the operations? Would the service be down?
Thank you.
client pay for storage he use, for bandwidth he download from storj, amount of segments.
there is some amount limints, but they are negotiable with contract.
If you pay your bills no one tell you that you downloaded to many times.
it is only 3K ops per second. not looks like a huge values …
The node operators would react to the new load and add disks or ssds to keep up.
The node operators with the fasest response would respond to your requests quicker and get rewarded.
The slower, more efficient nodes, would slowly fade to the background
The number of IOPS that Storj’s storage service allows is not publicly known. Storj mainly focuses on providing decentralized cloud storage, which relies on the collective storage capacity and bandwidth of participating nodes.
Contact Storj: If you have specific IOPS requirements, I recommend contacting Storj directly for more information.
Storj uses erasure coding to distribute data across multiple nodes. This can affect IOPS performance as multiple nodes need to interact for read and write operations.
Storj offers different storage plans tailored to different needs. It’s possible that some plans offer higher IOPS performance than others.
I’m not sure how well that really holds up. Back in the day, it may have been true, but not anymore. Read performance isn’t that important, since egress pays nowhere near what it used to. Write performance also isn’t important, because the storage node software writes asynchronously now, allowing it to win races despite having very poor performance on the underlying disk.
Wasn’t storj always telling us to “use what we have” and not to buy new hardware for storj. 
I’ve got history of totally ignoring that i.e. 5 nodes on the same HDD vs 5 nodes on a separate disks - #33 by andrew2.hart
The customer may start receiving 429 Slow Down responses if they hit a rate limit. Understanding Storj Usage Limits - Storj Docs. Can you say more about your use case? 3000 ops/second over 10gb of storage is a lot different from 3000 ops/second over 10 pb of storage.
5Tb only.
8 billion operations is the total for 30 days, but it’s not uniform, sometimes it has a “normal” usage and there is a few peak usage of 30Mb/s of many small files.
It will fit in Storj?
Why not? Did you try?
On the storage node side, we node operators will likely not even notice the increase of traffic, it will be lost in noise.
On the satellite side, Storj has rate limits. If you stay within them, it’s going to be fine. If not, you can negotiate higher limits with their sales department.
Some time ago we had a 6-digit need for cloud infrastructure and considered Azure. They arranged a meeting with a “sales” engineer. I was asking some basic performance questions, like what’s the expected latency between a VM and underlying storage… the engineer said to “just try it”. I laughed in his face and we went elsewhere. I wasn’t going to spend a week learning how to run accurate benchmarks on their infrastructure, while trying to understand all possible sources of variability of performance that may be specific to their infrastructure.
Yes, I understand your concerns. We do not have test results for every possible combination. Thus suggestion to try it, because even for the same ISP and the same speed it wouldn’t be the same as for the oldest one.