LakeEffect Research Cloud (On-Premise Cloud)¶

CCR hosts a subscription based on-premise cloud service affectionately known as "LakeEffect", which provides access to virtual servers, storage and networking that users setup and manage themselves. It is what is referred to as Infrastructure as a Service and is equivalent in functionality to Amazon's EC2 Cloud service.

Although high performance computing (HPC) is our largest and most widely used service, CCR recognizes that not all research requires number crunching on large scale computers (clusters). LakeEffect, was put into service to provide flexible infrastructure to support research computing such as webservers, databases, and long-running services that are not suited for HPC clusters. This along with our HPC platform allows us to provide a complete set of tools to support all the different research groups and collaborations throughout UB.

LakeEffect Quick Start¶

1. Details on Pricing and Subscription packages is here
2. Request access through the coldfront allocations management portal
3. Once you have an active allocation, you can access the cloud through the Dashboard
4. Documentation on using LakeEffect Cloud and Openstack:
   • Using LakeEffect Cloud
   • OpenStack Documentation

Features¶

Features of LakeEffect:

• Deploy virtual servers in minutes
• Storage volume creation/snapshot
• VPC Networking with access to public IP address space as well as the ability to create private networks
• Wide range of Machine Types available inclunding GPUs and IO Optimized Storage
• Various Linux operating systems are supported. NOTE: Windows images are not provided by CCR at this time
• Public Images of opensource Linux Distributions available as well as the ability for users to upload their own
• Load balancing of virtual resources across multiple physical machines
• Support for Floating IP addresses
• Easily manage access to VMs using Security Groups and Rules
• Built on Openstack Cloud Software
• Fault tolerant components to minimize risk of downtime due to underlying hardware failure

LakeEffect Resources¶

• 768 total CPU cores
• 8.8TB RAM (4GB per vCPU core)
• 1.7PB of object-oriented storage
• NVidia V100 GPUs
• NVME Ephemeral Storage
• Persistant Memory (PMEM) Storage

Subscriptions¶

In order to use LakeEffect Research cloud, you will need to purchase a subscription. We provide a package structure to make purchasing & tracking easier:

Package Name	vCPU	Storage	Other	Price
Starter	10 vCPU	1 TB	1 Floating IP	$500/year
Standard	20 vCPU	2 TB	2 Floating IP	$1000/year
Advanced	40 vCPU	4 TB	4 Floating IP , Fast-IO VMs	$2000/year
GPU Addon	8 vCPU	1 TB	1 Floating IP, Tesla V100 16GB	$880/year
Storage addon		1 TB	Block Storage	$55/TB/year

• Systems Administration Consulting: $63/hour

  • Must be purchased in 8 hour increments
  • Can be used in 30 minute chunks
  • Contact CCR Help for estimates on required time

How to Initiate a Purchase¶

To initiate the purchase of additional cloud storage or compute time, the PI or project owner should login to ColdFront and request an allocation change request on the group's Cloud Storage or Cloud Compute allocation. See here for instructions.

Use and Misuse¶

CCR Staff manages and maintains the Cloud platform, both the physical ifrastructure and openstack software. Users are expected to deploy and manage their cloud instances, including network security and protocols, installing and maintaining software, and running computational workflows. Users are “system administrators” of their cloud instance(s) and are responsible to ensure that UB Minimum Server Security and Hardening Standards are deployed.

Appropriate use cases:¶

• Academic courses (teach cloud computing, computer science, or HPC)
• Web portals for science gateways "Hubs" (JupyterHub, Shiny Apps)
• Running a research database/website application that interacts with the CCR Compute Clusters.
• Testing and development of software or products

Inappropriate use cases:¶

• Personal websites or websites unrelated to research
• Infrastructure that should be provided by IT Department
• Running workflows that can more efficiently be run through Slurm on our clusters
• Running VMs to Circumvent UB Server Security standards or IT Departmental regulations