Perumal Kuppuudaiyar's (Intel Lab Europe) talk at NC4 2016 was focussed on the implementation of test bed which had integrated with various state of the art software stacks on top of the heterogeneous resources to provide FT/HA clusters, fined grained resource management and containerised workload orchestration for HPC.
2. Topics
• Test bed Requirements
• Physical Layout
• Software Stacks
• Integration
• Use case Demo
◦ Interactive 3D Image rendering : Ray Tracing of a 3D Blender model using Intel’s Xeon Phi and
Embree Kernel.
08/04/2015 2
3. Requirements
3
• Managing Heterogeneous Resources
• Fault tolerance and High availability
• Support mixed workloads (MPI, Spark, Storm, etc.)
• Application portability
• Near bare metal performance
• Multi-node orchestration for parallel tasks
• Telemetry and workload characterisation
• …….
5. Software Stacks
5
Light weight portable
application container
Minimal Linux OS
- Systemd for running services,
- Etcd for clustering hosts together,
- Fleet for distribution across cluster
Mesos + Marathon +Zookeeper
Resource Management & Scheduling
Telemetry and Monitoring Systems
Weave creates a virtual network to connects Docker containers
deployed across multiple hosts and enables their automatic discovery
Container Cluster Manager
….
6. Integration
08/04/2015 6
StorageMemoryCPU GPUMIC NIC
fleet SDS dockeretcd SDN telemetry mesos
Marathon Kubernetes And other FWs...
PhysicalLayerDCOSFrameworks
Service Orchestration
(Bare Metals, Micro services (containers), KVM and Cloud formation Interface)
Service
Bare metal
(fleet units)
Dockerised
Applications Cloud formation
Clusters of
Containers
(POD)
7. Use Case – Ray Tracing
08/04/2015 TELECONFERENCE 7
• In computer graphics, ray tracing is a technique
for generating an image by tracing the path of
light through pixels in an image plane and
simulating the effects of its encounters with
virtual objects.
• The technique is capable of producing a very
high degree of visual realism, usually higher
than that of typical scanline rendering methods,
but at a greater computational cost.
8. Demo – 3D Image Interactive Rendering
8
Remote Client
Testbed Infrastructure
SSH and
X11 Display
Forwarding
VPN/SSH Tunnel
3D Designer create 3D contents
using CPU node
Interactive Rendering
using Xeon Phi & Embree Kernel
3D Designer create 3D contents
using CPU node
ServiceGateway-FW
Submit Job (2 CPU based blender instance &
1 Xeon Phi Embree pathtracer)
Orchestrator
CPU + Xeon Phi
CPU Node
CPU Node
CPU + GPU Node
CoreOS
Minimal Linux OS, Systemd for running services, Etcd for clustering hosts together, Fleet for distribution across cluster, Docker for running applications, Flanneld container networking
Mesos:
A distributed systems kernel
Mesos is built using the same principles as the Linux kernel, only at a different level of abstraction. The Mesos kernel runs on every machine and provides applications (e.g., Hadoop, Spark, Kafka, Elastic Search) with API’s for resource management and scheduling across entire datacenter and cloud environments.
Docker:
Container technology for the Linux platform, which leverages existing and well proven technologies such as control groups (cgroup), and LinuX Containers (LXC). Docker brings these technologies together and provides ease of setup, use and compelling efficiency
Linux containers include the application and all of its dependencies, but share the kernel with other containers. They run as an isolated process in userspace on the host operating system. They’re also not tied to any specific infrastructure – Docker containers run on any computer, on any infrastructure and in any cloud