Page 1 of 1

Docker/Containers

Posted: February 6th, 2019, 8:27 pm
by SWilson
Has anyone been building their apps using Docker for their heavy computational processes?  Or in general, has anyone been creating micro-services in containers for their programs for speed and scale?  

Re: Docker/Containers

Posted: February 10th, 2019, 2:01 pm
by finanzmaster
I am currently playing with docker. The business logic is based on the Monte-Carlo simulation.
And yes, my goal is to create a scalable (micro)service architecture. Since Monte-Carlo can be run in parallel, all I have to do is to put my container in Kubernetes and let it manage the balancing (at least I hope so).

Re: Docker/Containers

Posted: February 11th, 2019, 2:51 pm
by SWilson
I am currently playing with docker. The business logic is based on the Monte-Carlo simulation.
And yes, my goal is to create a scalable (micro)service architecture. Since Monte-Carlo can be run in parallel, all I have to do is to put my container in Kubernetes and let it manage the balancing (at least I hope so).
Sounds like we're doing something very similar.  Do you mind me asking if you're using a cloud service?  Azure/AWS?

Re: Docker/Containers

Posted: February 12th, 2019, 12:12 pm
by finanzmaster
Azure. But it shall not matter which cloud (or bare metall) solution one uses. At least, in theory :)

Re: Docker/Containers

Posted: February 19th, 2019, 6:05 am
by rrtucci
Bayesforge is available on the Amazon and Tencent clouds
www.bayesforge.com
https://market.cloud.tencent.com/products/8513#
We also have a Discourse forum at
http://discourse.bayesforge.com

Re: Docker/Containers

Posted: February 23rd, 2019, 8:36 pm
by ISayMoo
What is the overhead of this abstraction?

Re: Docker/Containers

Posted: February 25th, 2019, 3:07 pm
by SWilson
Thanks rrtucci,.  Our goal is to horizontally scale some of our functions so as far as actually stat libraries and functions we're ok.  Bayesforge appears free on the AWS marketplace so you pay for just the EC2 instance.  Been experimenting..... using containers and an orchestrator (kubernetes) is good if you're creating an entire ecosystem of micro services, but I've been playing around with AWS Lambda as well since all I'm trying to do is move some of the heavy computation off of one resource and scale into another (serverless) architecture.  

Re: Docker/Containers

Posted: February 25th, 2019, 5:53 pm
by rrtucci
Thanks rrtucci,.  Our goal is to horizontally scale some of our functions so as far as actually stat libraries and functions we're ok.  Bayesforge appears free on the AWS marketplace so you pay for just the EC2 instance.  Been experimenting..... using containers and an orchestrator (kubernetes) is good if you're creating an entire ecosystem of micro services, but I've been playing around with AWS Lambda as well since all I'm trying to do is move some of the heavy computation off of one resource and scale into another (serverless) architecture.  
Thanks to you SWilson. Bayesforge was written completely by my cofouder Henning Dekant(originally German, but now lives in Toronto) except that my other co-founder Dr. Tao Yin (he lives in Shenzhen) submitted Bayesforge to the Tencent cloud and jumped through the hurdles to get it approved there. Having a presence in both Canada (software exports less regulated than in USA) and China is an essential part of our strategy. I am mostly responsible for the quantum side of the business, but the beauty of Bayesforge is  that it can also be very useful to Data scientists that have ZERO interest in Quantum Computing. I see Bayesforge as a competitor to Google's Colab. Do we want MAFGA (portmanteau of Mafia and MAGA and also an acronym for Microsoft, Apple, Facebook, Google,  Amazon) to control everything in the world? 

We would totally love to collaborate with you. We currently have no funding but big aspirations. I think Bayesforge is a billion dollar opportunity. Email me at tucci@ar-tiste.com if you are interested in collaborating with us. Henning knows Linux inside out