SONM opens Decentralized
Fog Computing Challenge

Total prize fund of $200 000

In the run-up to the launch of the first decentralized Fog computing platform, SONM invites developers and tech enthusiasts from all over the world to take part in a multidisciplinary fog computing challenge.

SONM is a decentralized fog computing platform and a marketplace for computing power for real world applications: from CGI rendering to machine learning, and scientific calculations.

THE CHALLENGE

This online-only challenge is made up of two stages and open to anyone ambitious enough to test their skills in solving the most cutting-edge tasks for the one of the hottest projects in the industry.

Team

You can work alone, or with a team. One person could win as many nominations as they can. The final decision will be made by Advisors based on the code you send to us.

Stage 1

April 25, 2018 — May 25, 2018

This is about development on Docker Containers and is open to all to enter. There will be 14 tasks to choose from, and you will work with SONM Advisors on the task you picked.

Stage 2

May 29, 2018 — June 29, 2018

Deployment on SONM test network with the new API (v2). This stage is only for those who have been successful in the first challenge and have been invited to participate in the second round.

As part of the second stage, you will have access to the SONM main net and will be provided with an amount of SNM tokens for internal exchange. You’ll also have the unique opportunity of being the one of the first to run live applications – as well as grow your presence - on our platform.

FAQ

  • This challenge is open to all, so if you prefer to work alone, that’s fine. Or, if you want to enter as a team, that’s fine also. One person has the potential to win as many nominations as they want.

    The task prize money varies according to task difficulty, so it is advised you choose the right task according to your skillset.

  • This challenge is an amazing opportunity to be part of SONM, get the early access to SONM platform, make a real impact on the most powerful supercomputer and be rewarded for it.

  • There is no admission fee. Anybody with tech know-how is able to enter.

    Step one: simply choose the topic you want to work on and submit the

    Step two: your Advisor will contact you directly via email.

  • Stage one: Development on Docker containers out of SONM network.

    Deadline for submissions: May 25, 2018

    Stage two: Deployment on SONM with new API (v2).

    Deadline for submissions: June 29, 2018

  • To submit your work for Stage one:

    Email your solution to the relevant Advisor via email. If everything is alright, your Advisor will give you the invitation to the next stage.

    Note: Only those who will successfully pass the first stage, will be invited to participate in the second stage.

    By the end of second stage the final decision will be made by Advisors based on the code you send to us. Winners names will be revealed on July, 3rd.

    In case there will be no solution submitted that fit the necessary quality criteria and level of elaboration, SONM retains the right not to award the prize.

    SONM retains the right, at our discretion, to encourage a team that has developed a good but not the best solution. For that purpose SONM is reserving an additional fund of $25 000.

SONM Fog Computing Challenge

SONM Core

  • Provide a solution description and a prototype to prevent suppliers from intruding SONM tasks.

    One of the biggest challenges for SONM is enforcing security on running tasks. The supplier has full access to the PC with running tasks, so they can get all task data and code - and even emulate tasks, thereby providing false results to the customer.

    To prevent an attack, one should guarantee that a certain OS image with unpatched Linux kernel is started with enforced high-security options in that kernel. Not only that, but the image should also be able to cryptographically prove the fact that it is executed in trusted mode. The solution must be compatible with most of the existing mining hardware, with additional hardware costs not exceeding 50$/PC.

    The solution should also be accompanied with a prototype demonstrating the crucial technologies of the solution.

    Advisor

    Select Advisor to apply for!
  • Develop the automatic trading software to maximize SONM supplier’s income.

    The software is/has to run on SONM’s supplier’s side 24x7. It also has to use SONM’s public API to monitor the market, check available resources, create Sell orders, manage deals.

    Advisor

    Select Advisor to apply for!
  • Develop the automatic trading software to run cloud mining on SONM.

    The software is to run on SONM’s customer side 24x7. It has to analyze the cryptocurrency market to evaluate mining incomes. It also has to use SONM’s public API to monitor the market, create Buy orders, start mining tasks and monitor deals.

    The successful software will enable clients to earn more money by running cryptocurrency tasks.

    Advisor

    Select Advisor to apply for!

PaaS

  • Deploy an object storage, backed by SONM resources.

    Choose an open source object storage and make it run using SONM resources. The users should be able to use the object storage as a cloud service, using S3 or Swift interface to customers. The service should use SONM public API to buy resources on the SONM market and deploy tasks for slave nodes.

    The successful software should provide acceptable performance using SONM distributed networking. The software should be able to scale well while stored data is growing and should successfully resist node failure.

    Advisor

    Select Advisor to apply for!
  • Deploy a network file storage, backed by SONM resources.

    Choose an open source network storage and make it run using SONM resources. The service should be able to provide NAS to customers. The service should use SONM public API to buy resources on the SONM market and deploy tasks for slave nodes.

    The successful software should provide acceptable performance using SONM distributed networking. The software should be able to scale well while stored data is growing and should successfully resist node failure.

    Advisor

    Select Advisor to apply for!
  • Deploy a NoSQL KV storage, backed by SONM resources.

    Choose an open source KV-storage and make it run using SONM resources. The users should be able to use Redis, MongoDB or other well-known APIs to manage data. The service should use SONM public API to buy resources on the SONM market and deploy tasks for slave nodes.

    The successful software should provide acceptable performance using SONM distributed networking. The software should be able to scale well while stored data is growing and should successfully resist node failure.

    Advisor

    Select Advisor to apply for!
  • Deploy a RDBMS, backed by SONM resources.

    Choose an open source RDBMS and make it run using SONM resources. The users should be able to use Postgress, JDBC or other well-known API to manage data. The service should use SONM public API to buy resources on the SONM market and deploy tasks for slave nodes.

    The successful software should provide acceptable performance using SONM distributed networking. The software should be able to scale well while stored data is growing and should successfully resist node failure.

    Advisor

    Select Advisor to apply for!

SaaS

  • Run Golem “mining” on SONM.

    Golem is a distributed supercomputer project based on cryptocurrency. It’s capable of running grid computing tasks, thus generating revenues.

    Create docker image and SONM task to run Provider Golem node by customer. The node should be able to execute Golem tasks in Golem network and earn tokens.

    Advisor

    Select Advisor to apply for!
  • Run iEx.ec “mining” on SONM.

    iEx.ec is a distributed supercomputer project based on cryptocurrency. It’s capable of running grid computing tasks, thus generating revenues to spare resource owners.

    Create docker image and SONM task spec to run iExec node by customer. The node should be able to execute iExec tasks in iExec network and earn tokens.

    Advisor

    Select Advisor to apply for!
  • Deploy CGI rendering farm, backed by SONM resources.

    Rendering farm is a system that manages CGI rendering. It can render multiple scenes simultaneously using existing hardware. It supports queueing and CGI modelling tools integration. At the core of rendering farm is a render farm manager software, which manages users, their rendering tasks and computing instances.

    Choose rendering engine software (open source or commercial) and prepare docker image and SONM task spec. Choose rendering farm software manager (open source or commercial) and deploy a service. Configure the service to use SONM public API to buy resources on the SONM market and deploy tasks in SONM for render nodes.

    The successful software should provide acceptable performance using SONM distributed networking. The software should be able to scale well while task queue is growing and should successfully resist node failure.

    Advisor

    Select Advisor to apply for!
  • Deploy distributed ML service backed by SONM resources.

    TensorFlow is one of the most popular ML toolkits. It enables neural network training and applying trained models to process data.

    Develop a service that will provide neural network distributed training. Configure the service to use SONM public API to buy resources on the SONM market, and deploy tasks in SONM for render nodes. Choose a demo model and training data at least 10GB in size. Train a model and apply it to some demo data.

    The successful software should provide acceptable performance using SONM distributed networking. The software should be able to scale to increase neural network training speed. The software should successfully resist node failure.

    Advisor

    Select Advisor to apply for!
  • Deploy Apache Spark cluster backed by SONM resources.

    Apache Spark is a framework that enables efficient big data processing using SQL-like queries and Scala/Java/Python/R distributed programs.

    Deploy Spark Master as a service. Prepare docker image and SONM task spec for Spark slaves. Configure the service to use SONM public API to buy resources on the SONM market and deploy tasks for render nodes in SONM.

    The successful software should provide acceptable performance using SONM distributed networking. The software should be able to scale well while stored data is growing and should successfully resist node failure.

    Advisor

    Select Advisor to apply for!
  • Deploy content delivery network backed by SONM resources.

    Content delivery network software enables fast content delivery to the client. Individual web servers face performance and network bandwidth limitations. CDN can automatically copy content to other servers, so that traffic can be split between multiple servers all over the world, enabling global load-balancing capability.

    Choose an open-source CDN software, that is capable of serving static web files or stream audio/video. Prepare docker image and SONM task spec. Develop load-balancing and discovery software that will use SONM public API to buy resources on the SONM market, and deploy tasks for CDN nodes in SONM.

    Advisor

    Select Advisor to apply for!

Games

  • Develop an engine for a classic Game of Life, that will be able to work using combined power of multiple computing nodes from SONM IaaS.

    Requirements:

    • Game field (space) must be distributed into sectors. Each sector must be processed by a different computing node.
    • Computing nodes will be hosted in SONM.
    • Computing nodes may be of different performance. Game engine must be capable of leveraging computing nodes of unequal power.
    • Computing nodes may be connected by networks with different latency and throughput. Game engine must be capable of leveraging network effects.
    • For a spectator inside the Game universe the game space must look continuous, sector boundaries must look transparent.

    Problem: Game of Life has “stages”, each stage synchronous execution and continuous space. Cellular automata is synchronous (stages) and can not work separately.

    Solution: one global stage may be split into separate zones, that are not linked together.

    Advisor

    Select Advisor to apply for!
Official rules

Social media

Find insights, discuss and contribute

SONM Official Site