Elevated supplier rewards program
1. Check that only spot deals are currently open;
2. Run the script (you need to replace MASTER_ADDRESS with your master address). This script will close the current spot deals and update all platform components:
sudo service sonm-optimus stop && sonmcli worker ask-plan purge && sudo bash -c “$(curl -s https://raw.githubusercontent.com/sonm-io/autodeploy/master/sonm-auto-deploy-supplier.sh)” -s MASTER_ADDRESS
At the moment, we have attracted enough 1080 Ti GPUs and temporarily suspended the admission of new suppliers to this program. As soon as we need more 1080 Ti, the attraction of new suppliers will be resumed.
Suppliers who have already joined the program will receive elevated rewards for at least 3 months (starting from April 15), as we promised.
A month ago we announced turning Crypto-IaaS into a full-featured Cloud platform for startups and individuals. The key idea is to build a set of tools on the top of SONM infrastructure as an abstract layer that hides the complexity involved in provisioning and managing bare instances for various GPU-related computations.
At the initial stage, we focused on 2 services:
- ML service for building, training and deploying machine learning models;
- Video Rendering service designed to help freelancers and small studios to render scenes in just a few clicks.
These are the first steps in building our two-sided platform – we have collected instances on the supplier side of the platform and now we are building ‘one-click’ services on the customer side. You can look at these services as ‘Uber for rendering’ and ‘Uber for machine learning’.
We’ve already launched Rendering service in closed Alpha and we hope to make it public soon. Machine Learning service will be launched around the beginning of summer.
However, both to launch these services and to implement our further plans, we need not only connect many devices to the SONM platform but also to balance the healthy proportion of various configurations.
- Some rendering tasks require 5-6 Gb vRAM and can be run on a wide range of Nvidia GPUs from 1060 to 1080.
- Other tasks require more vRAM (in the range of 8 to 11Gb) and can be run only on high-end GPUs, like 1080 Ti.
- But some tasks require even more than 11Gb vRAM, and need to be run either on non-consumer GPUs (Tesla) or on a CPU.
If we plunge into the jungle of Machine Learning tasks, the situation will be even more complicated and we will need even more configurations of GPU rigs and Servers.
Currently, most of the thousands of devices connected to the SONM platform are mining rigs. This naturally leads to a large number of GPUs with 8Gb vRAM in any possible configurations – we have more than enough of them to start the Rendering service.
But the GPUs with 11Gb vRAM are currently a bit lacking, and we would like both to increase their number and to get a greater variety of configurations, with various amount of RAM, disks and types of processors. Moreover, in the near future, we may need enterprise servers or even home computers of various configurations.
Therefore, we decided to launch an elevated rewards program and stimulate the connection of those devices that are currently most needed for Rendering and Machine Learning services.
Let’s take a deeper look at what will drive the supply in the coming months.
This is a long-term program that will continue for at least a year. We will monitor the balance of various types of devices on our platform and announce a reward increase for those types of GPUs or Servers that will be in high demand among the Rendering or Machine Learning customers.
- If, as now, we see a shortage of 1080 Ti, then we’ll announce higher rewards for this type of GPUs.
- If we miss 1070, then we’ll separately announce higher rewards for this type of GPUs.
- If we need enterprise servers with CPU and storage, then we’ll announce elevated rewards for them.
Each time we’ll announce the minimum period (how many months the elevated rewards will continue) and how large this reward will be.
Our primary focus over the next few months is to keep the supply at least 2 times the current demand for each GPU type. We don’t want to face the lack of equipment our customers need for rendering or machine learning.
Technically, this program will work similar to the current guaranteed devices utilization. When you connect your rig with 1080 Ti to the SONM platform, our algorithms will automatically utilize it 24/7 by switching from performing useful computations to cryptocurrencies mining and vice versa.
No matter what tasks your device performed – you are guaranteed to get an elevated reward for every hour your 1080 Ti was connected to the SONM platform.
As mentioned above, our first step will be the introduction of elevated rewards for the Nvidia 1080 Ti / 2080 Ti, in order to increase the number of these devices before starting the Rendering service.
- x86-64 architecture;
- CPU – Celeron or higher (at least 2 physical cores, 2 Ghz);
- GPU – Nvidia 1080 Ti or Nvidia 2080 Ti. The device must have at least 1, but not more than 8 GPUs. The reward will depend on the number of connected GPUs (you get ~$2 per day for each GPU).
- RAM – 16Gb or higher. RAM over 16Gb will not increase your reward but is welcome.
- Disk – 60Gb or higher. Disk size over 60 Gb will not increase your reward but is also welcome.
- Internet channel – a stable wired Internet connection, with a speed of at least 10 Mbit/s.
Any device that meets these requirements will be automatically included in the program when connected to the SONM platform. For this you need:
- Make sure that your device meets the requirements, upgrade it (disk, memory, GPU) if necessary.
- Follow the guide: “How to become a supplier“
If your device meets the described requirements and is already connected to the SONM platform, then nothing needs to be done – we will include your device in the program automatically.
Elevated rewards for 1080 Ti and 2080 Ti start from April 15th, 12:00 UTC, and will last for at least 3 months.
We’ll monitor the balance of devices and may extend this period up to 1 year. We’ll also keep you informed about any changes or elevated rewards for other types of devices.
Program funding source
We recently published a half-year report summarizing the period of the SONM platform’s existence since its launch in July 2018. The metrics and feedback from the first 100+ customers look very optimistic.
We are also enthusiastic about the upcoming launch of Rendering and Machine Learning services – for instance, a number of video production studios are already testing the closed alpha version of Rendering service and sharing very positive feedback.
In this regard, the SONM team decided to invest their own tokens in this program. The team’s token pool is 10% of total supply (= team holds 44,400,000 SNM), and most of them will be directed to this program. The remaining part will be used to subsidize the planned Masternodes during the first years of their work.