SONM is a based distributed computing marketplace, built on the ethereum network.

While similar to Golem in its aims, SONM has a wider range of features and does not include a platform for developers to build new applications on top of. The latter means running the system is more straightforward as the network does not require complicated mechanisms for ensuring applications do not include malicious code. The former features include but are not limited to a computing market place and automated mining matching.

SONM’s computing marketplace functions as a regular exchange. Providers of computing power will automatically be matched with the customers offering the best prices. This framework is also very useful for miners. Miners participating in a proof of work network like bitcoin or Ethereum often go through prolonged periods of minimum profits as there are fewer transactions to process and network fees decrease when traffic is low on these networks. SONM’s platform will allow these miners to switch to mining the most profitable coins at any given point in the network. Moreover, miners could also use their hardware to provide computing power to the SONM network. The level of profitability will be the determining factor in all of these cases.

The team provides many use cases for their network. Below are a few:

Software as a service/AI engineers:

SONM promises to provide unlimited resources for both neural network training and application, competitive pricing (eg. a PC with Nvidia Geforce 1070 GPU card will cost about 0,12USD/hour, comparing to minimal 0,4USD/hour on the market), strict security for customer’s data from data service centers, compatibility with machine learning software and integration with cloud services for data storage (eg. Amazon, S3 Dropbox etc). However, there are limitations for those wishing to use SONM for machine learning. Because the system is distributed all around the world, there will be poor internet connectivity in some cases. Because of the large amounts of data required for machine learning, this could be a problem (for machine learning (ML) systems like Leela Go, this should not be an issue.  This limitation is mediated by the fact that connections from dedicated partners will be as fast a local networks and ML training data is usually hosted on external internet resources (cloud storage and online services).

Big Data

SONM can provide huge computing resources on demand (disks, CPUs, GPUs) and low competitive prices. However, distributed computing can also be a problem in this regard.

Cloud Service Providers as Customers

Cloud providers usually need to setup additional hardware to backup user data. Instead of using the extra hardware, some users do so by renting extra space on cloud services like amazon or google. Renting such space on SONM’s network will be much cheaper and the latency issue won’t matter as much in this case. Additionally, cloud service providers could simply buy space on SONM’s network to store the users data. The hardware and cost would be cheaper.

Another example is starting GPU resources to the cloud provider customers. GPU is not common hardware for cloud provider infrastructure. Still this market is evolving and some cloud providers are willing to suggest new options to their customers. In this case SONM can provide unlimited GPU resources to provider and decrease market entrance costs and risks.

Limitations to these approaches include potential problems with connectivity on a distributed computing system. This can lead to poor calculation performance.  Moreover, most providers on the SONM network will not have the same high level service agreements that usual cloud based services do. This means that node failure is a possibility. Applications must make sure to be set up in a way that migrates to different nodes in case of node failure.

Cloud Service Provider as customers

Cloud service providers often have excess resources in case customers need them in the future. These providers can mediate these costs by renting spare resources to the SONM network to generate extra income. For this purpose, SONM provides applications for these providers to spot profitable opportunities to provide computing resources.

Other use cases include computing intensive projects, Computer graphic studios, render farms, and render software developers.


SONM is in the first generation of distributed computing blockchain projects. There are many limitations to the system that may not justify use over a regular computing platform. This is especially true for projects that require a very reliable cloud computing infrastructure. The company will need to seriously update their infrastructure to compete with the next generation of blockchain based distributed computing systems that provide better incentives to create wider and more concentrated distributed computing networks. However, their first mover advantage and excess resources due to the current blockchain ecosystem makes them a project worth monitoring.



Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s