Help:Architecture

Jump to: navigation, search

Deployment

  1. With the Kogence Individual and Team Subscriptions, you would log into kogence.com container cloud HPC PaaS platform. You can keep all your content (you models, data, software containers, discussion boards, wiki pages etc) private to your or share with your collaborators. If you want you can selectively open any of these components to public. On the other hand, with the Kogence Enterprise Subscription, the entire Kogence HPC Grand Central App is privately deployed specifically for your organization.
  2. Privately hosted platform includes the private version controlled git-LFS repository of your enterprise scientific models on a git-LFS server.
  3. It also includes, private version controlled container repository of your enterprise scientific software applications.
  4. License manager also runs within your private deployment.
  5. Your employees can selectively open any content to everyone in your organization or keep private to themselves or share with their collaborators within your organization. Your organization would assign a platform admin who will have unrestricted access to all content.
Kogence Platform Architecture v3.png

Architecture

  1. On Kogence platform, all application software, simulators and solvers are deployed in independent individual Docker containers. Each software application is provisioned as an HPC microservice running on HPC cluster hardware you requested in the Cluster tab of your model. You can build complex simulation workflows combining multiple containers using the Stack tab of your model. Kogence' kComposition technology automatically composes the microservices into full software stack without any effort on your part. Solver binaries in one of the containers can be called from any of the other containers selected in the stack. You do not need to write any composition scripts. Containers remain completely general and fully independent of each other. Check kCompostion for more details.
  2. Kogence HPC Grand Central App automatically orchestrates the most optimized cluster on-the-fly depending on the workload you submit. As your simulation scripts execute, workload may need different type (e.g. CPU vs GPU) and different number of nodes in your cluster. Kogence automatically creates and destroys nodes in full heterogeneous cluster depending on your workload needs. All nodes in a cluster are connected to a high speed file system which is destroyed on completion of your workload. Kogence HPC Grand Central App is dynamic and multi-cloud. Requested compute resources comes up on the infrastructure that is most cost effective and is most appropriate for the workload submitted. Check kScaling for more details.
  3. If a node in your cluster fails, ceases to be available or if you application needs more hardware resources than what can be supported by the node, Kogence HPC Grand Central App will ensure that your workload does not fail but is migrated to a new node automatically added to the cluster while preserving the state of all running processes. Check kMigration for more details.

Security

  1. Entire Kogence HPC Grand Central App is behind your corporate firewall. App will only be accessible from IP addresses you specify. Entire infrastructure is contained in a Virtual Private Cloud (VPC). VPC is connected to your enterprise Virtual Private Network (VPN).
  2. Your enterprise model repository is encrypted with private encryption keys that only you own. Kogence does not have access to keys and if you loose keys it will not be possible to recover models.
  3. All communication between Kogence HPC Grand Central App, HPC compute nodes and user's client machine (onprem laptop, workstation), including the remote interactive graphics, is channeled under industry standard secure tunnels (SSL/TLS and ssh).
  4. Client side authentication enforcement can be opted for in addition to secure communication tunnels.
  5. Once compute nodes terminate your data volumes are deleted as well. Data is only stored in encrypted git repositories on deployed git-LF server. For additional security you can opt to use an onprem git server. In that case, once compute node terminate, no data would be left on cloud.
  6. Our corporate security policies: https://kogence.com/app/landing/security

Contact Us

Please contact us to discuss further.