SNOs operating multiple nodes and using the multinode dashboard should be taken into account also.
You should mention that on the github page.
I don’t have a Github account. Otherwise I would.
No, that is not the point. The point is that if
A: You need high performance no matter the cost, you are unable to get it from STORJ (according to @Knowledge numbers, it is 16x slower)
B: You have a VPS on DigitalOcean or OVH or any other provider, their “local” S3 bucket will offer a better price/performance ratio.
B is the reason for this post Update Proposal for Storage Node Operators - #767 by IsThisOn that everyone will ignore and wonder why customers don’t switch 
In my dreams, STORJ runs these benchmarks and promote ads like “Hey DigitalOcan users, did you know that our STORJ S3 bucket offers twice the performance for half the price? Try it today with this 10% promocode”
2 Likes
Wouldn’t local storage always be faster? Fastest would be the drives inside the server, slightly slower is a file server in the same rack or datacenter and slowest is anything outside the datacenter where your VM is.
IMO Storj and similar services are only useful if the data is accessed from multiple different locations. Otherwise, just use whatever your datacenter has, since it will be fastest and/or cheapest.
2 Likes
Gosh,
Currently I’m managing a little over 130TB of StorJ storage in different data centers around the globe which has been mostly a “hobby” for the last 5 years. But giving me a 50 to 75% haircut on earnings will definitely have me looking at changing to some other sort of gratification. At the current level there was at least some sort of reward. Please don’t trim more than 25% off especially for the older nodes…
some sort of conspirology: looks like high incoming \ outgoing “test data” last month is generation of extra FEC to mitigate massive SNO quit on payouts decrase ?
1 Like
Test data is on test satellites. That hasn’t changed. There are some challenges and fixes going in to manage the free tier better from abuse. Once that is in place, we’ll get a better understanding of normal data flow. I suspect some data will fall off after a time as some terms of use abusers are removed from the system.
7 Likes
Sure, local storage would be faster, but also more expensive. For DigitalOcean 1TB costs 100$ while spaces objects (S3) are around 25$.
The problem is, we narrow down the user groups more and more. If we would listen to this forum
- We don’t want or can’t the high performance users, simply because we can’t compete because STORJ does not offer compute
- We don’t want or can’t the backblaze users, because SNOs think STORJ is better and worth 20$/TB and because that is what nodes need to buy new hardware.
- We don’t want or can’t gov, big corp or healthcare because regulations
- We want the users that store highly read intensive files. Write once, read thousands of times all over the world, and with higher speed requirements so we can justify higher prices than the competition
Who are these users? Test data?
for GatewayMT in some region (I do not have any of these numbers, but I believe it’s 100Gbps).
If you want to have speed - you must use a native integration where is possible. Even if your software have only S3 support, you may run your own GatewayST and integrate with it, increase parallelism and you will saturate literally any bandwidth.
And all usage from the customers are welcome, include performant users, backblaze users and govs (we have a geofencing feature and with latest information, we can comply with GDPR).
3 Likes
I think the problem is that the service is one-size-fits-all, which of course, if it fits everybody, then it fits nobody.
Other companies have different services based on use case. Backblaze offers cheap storage for backups. It probably is not particularly fast, but good enough for backups and, importantly, cheap.
On the other hand, if I want a CDN that can do hundreds of requests per second, I’ll need to use some other service that is more expensive, but offers the speed.
And, of course, everything in between.
Storj offers one type of service that tries to do everything - be as fast as the fastest competitors, but also be as cheap as the cheapest (slower) competitors, then we get into this weird situation where we need the performance but there is not enough money.
I can compare it to VDS offers from various companies. If I want to store a lot of data and performance is not as important, I would prefer a cheaper service that uses hard drives. I understand that SSDs are better and NVMe SSDs are even better than that, but it won’t matter to me if the IO load is 2% or 0.2%, but the price would matter.
On the other hand, if I need the performance, I will pay more to get a VDS on a host with SSDs.
1 Like
We have been using Backblaze B2 for about two months now and we have image backups using Veeam tied to it both from their Service Provider Console and also using Synology Cloud Sync. Speed has been surprisingly good - considering we are doing image backups. Our limitation has been uploading large files using the Synology Cloud Sync app - for some of our larger storage systems we will need to use the native Backblaze app and do file level backups only. But I have no complaints on it currently.
1 Like
The user does not have to do anything. We are selling the product to them.
Again, of course all customers are welcome, we are selling a product. The problem is not that we don’t want them, the problem is that they don’t want us.
@Pentium100 Totally agree with all you wrote. I just hope for the ICO money to run out soon, so we finally see some real world results. I still think this project is extremely interesting. In the end it is a delicate balance between all these numbers.
Here is how I think it went down so far:
STORJ know who their customers are and the have a sales team. For them it is very clear and they also clearly stated that rising prices is not an option and would scare away customers. So the looked at the running costs and calculated how much the can pay the nodes to be profitable and came up with these proposed numbers. SNO reacted very harsh and threatened to quit. STORJ was surprised by the backlash. They backed down a little bit and wondered if it is only a loud minority. So they put up a google doc survey. My guess is that they looked at numbers and realized, that they are in deep in the dirt.
One option would be to upsell but their sales team does not think this is going to work.
The other option is to just hope their will be enough nodes around with the new proposed pricing.
Another huge risk for them is the ICO money is volatile. It is 30cents now but with Binance stopping withdrawals today, this could go down fast. So maybe they don’t have until the end of the year and have to come up with a solution faster.
That’s a weird request you have there for a start-up company to quickly lose its reserve cash so it can go out of business. You do understand that start-ups operate in their early years at a loss, right? And without cash reserves, start ups fail. You can’t expect the company to suddenly be profitable. That’d be nice if it was, but that’s not how it typically works. I’d like to find a pile of gold in my backyard, but it’s unlikely to happen.
6 Likes
The problem is that at least I can’t really think of any way for Storj to make different services for different prices other than selling the edge services separately.
Maybe have different classes of nodes - slow and fast. Slow nodes are slower but are paid less, great for backups and SMR drives etc. Though with 30 nodes for one segment it may not matter that much and there would be a problem proving that the node belongs in the “fast” class.
Just shuffling the prices around? Backups pay less for TBm, more for egress, CDNs pay more for TBm, less for egress or something.
Because we are selling some kind of hybrid that does not really fit the needs of anyone. On average it fits everyone’s average needs, but it does not ft individual needs.
Also, Storj is insistent to only sell to developers and not end users, at least it looks that way.
Backblaze offers a backup program that uses their service, but if I wanted to back my files up on Storj, I would either have to look for a third-party program (and have lots of fun if something does not work right) or write my own scripts that use the uplink.
3 Likes
What software does Amazon S3 offer? I only know about the CLI.
Isn’t it good to let everyone focus on what they know best? Storj knows how to store data, someone else knows how to create a nice backup software. Backup software also needs more features than just moving files from A to B and back again.
5 Likes
I’m not a salesperson but I think that the initial target of Storj is wrong. The Storj network has huge potential as an AWS alternative but it may still be too early to focus on that.
What I see here is a massive opportunity that is being put aside: Web3. Decentralisation is still very crypto-related in people’s minds and when trying to sell Storj as an AWS alternative for anybody outside Web3, you first have to explain why decentralisation can be beneficial to them before exposing the functionalities of the product. You don’t have to explain this to people who already are involved in Web3 because they know what it means to them.
Sure FreeNAS and other partnerships are great milestones for Storj but I’m extremely surprised about not seeing Storj being the n°1 hosting platform for NFT media and metadata as an example. Web3 people and projects need to learn more about Storj and a huge part of this has to go through Storj going ahead and making sure that these people understand that launching an NFT collection and hosting images on Google Cloud is nonsense.
Traditional and established storage solutions partnerships and integrations are great. But before this make sure to bootstrap sales by putting the product in front of already convinced people’s eyes.
2 Likes
Amazon is big and it can get away with it. However, Amazon offers AWS and if I decide to run my VMs on there, it seems logical to store my data on Amazon as well, fewer compatibility problems, fewer pointing fingers in case something goes wrong and fewer different bills to remember to pay. As I see, if you use Amazon CloudFront CDN, then data transfered from S3 to that is free. And you probably pay less if you access the S3 stored data from a VM running in AWS.
Amazon also has service tiers - backups can use a slower, cheaper service while frequently-accessed data has to use a more expensive service.
There are a few problems with this, at least for a small company like Storj.
First, it may increase the price. Let’s say someone wants to back his files to to Storj. To do that, he needs to buy compatible backup software (there is no reason why it has to be free) and then pay Storj for the storage. This adds up to probably higher price than paying Backblaze for storage and getting the program for free.
Another problem is support. What if something goes wrong. It could be a problem with Storj or a problem with the backup program. The support of each may start pointing fingers at each other. I have seen this quite a few times, not with Storj, but in general.
A third problem is how to convince the developer to add Storj compatibility or make a program purely for Storj. Clients won’t really demand it because they don’t know what Storj is since Storj does not market to end-users. In comparison, pretty much everyone knows what Amazon is.
The developer probably does not have a reason to add it, unless he gets kickbacks (which would probably increase the price) and if he does add it, he will most likely just use the S3 gateway because it’s easier (if the program already supports S3 protocol), but it is more expensive for Storj.
Even if the developer adds it, now you still have to convince the end-user to use it instead of the other options which are probably supported by the same program. How? Leave that up to the developer to convince the user? Why would he do that? He already sold the program, it probably supports other storage services that the user uses.
For example, The “Hyper Backup” program on Synology NAS (why Synology? I have access to one) can back files up to multiple cloud services - Synology themselves, Dropbox, Azure, Google Drive, Rackspace, some others and even S3. So, in theory any Synology user can back his files up to Storj right now (using the expensive-to-Storj gateway of course, but whatever). How many would do it? How many know Storj even exists? Why should they choose Storj instead of the others? By the way, a random NAS user won’t need 5Gbps transfer speed, so he will either choose the cheapest option, the easiest option or one he knows best (“oh, it’s Google” or “well, my NAS is Synology, might as well use their cloud”).
Is there even a backup program for Windows that uses Storj? Other than the programs that only support S3 and need the gateway which removes on of the advertised features of Storj - end-to-end encryption.
2 Likes
First of all the person has to know about storj. If someone gets to know about storj I expect it not to be a completely tech illiterate person. Because if they are, they do their backups on hdds and DVDs/BluRays, since that’s easy. I don’t think many non tech people will create S3 Accounts or any other accounts. What kind of data do most people have? Mainly Images. Google and Amazon Prime offer “free” services for photo backup.
Tech people won’t have issues to find a software that works with storj. Via gateway of course, but that may also have been storj marketings fault, since they promoted it so much.
Having another software to care for splits the developer time. Fixes, features, support has now to be done on 2 products. If the free account is not reduced then there would still be no income. If I look at my important data then it’s less than 100GB and most of that is photos. I don’t know how it is for other people.
I still believe the focus should be on B2B and developers for storj. Consumers don’t bring much money, only need more support.
2 Likes
Seagate launched Lyve cloud with 9$/TB stored and free engress. So… where do we stand with 1.5$/TB?