Emind’s Architecture forEnterprise & AWS
IntegrationLahav Savir, Architect & CEO
Emind - Your Cloud Experts
Reduce wait time in queues.
Access to endless computation resources.
Easy scale-out of existing computational pipelines.
Collaboration of projects with external research institutions.
Enable access to the rich variety of cloud services.
Combined Self & Managed Service.
Reduce CAPEX.
Goals
Global ReferencesHarvard University -
http://huit.harvard.edu/services/harvard-cloud-services
University of Chicago - https://aws.amazon.com/solutions/case-studies/university-of-chicago/
University of New York - https://aws.amazon.com/solutions/case-studies/new-york-university/
Scope
Advanced Computing
Data PlatformsHPC ClustersWeb Hosting
Integrations
Billing and AccountingNetworkingSecurityComplianceIdentityApplications stackMonitoringUsage analyticsAutomation
Managed Services Self Service
Billing & AccountingAllow “Pre-Paid / Budget” payment
Controlled budget
Pay per use
Automated alerts and actions
Integration of the AWS billing with the organization’s internal billing
Seamless ordering / allocating budget
Seamless usage / consumption information
IdentityRemote Resources:
AWS Console
AWS API
AWS Network
EC2 Instances
SSO (Single-Sign-On) integration with Organization’s Idp
Authentication
Role Inheritance
Network & ConnectivitySeamless experience - ONE Network
Extend organization’s network to AWSas an integrated data center
Full routable network
Redundant and ScaleDirect-Connect
Highly Available VPN
IPSec Tunnels
SecurityUnified Policy Enforcement
Controlled network
Monitored in/out traffic
Network IPS
Role based resource access
Host Protection
Vulnerability Scanning
Host based IDS and FIM
Events Correlation and Alerting
Comprehensive Audit
Application StackSeamless access to
organization’s tools libraryShared scripts and modules
Shared data sets
Predefined instancesIntegrated tools
Integrated authentication
HPC FrameworkStarCluster
http://star.mit.edu/cluster/
MonitoringCentral Monitoring
PlatformDynamic ProvisioningPredefined TemplatesIntegrated Data
SourcesCompute instances
AWS resources metrics
Usage Reports and AnalyticsAnalyse HPC
resources consumption
Periodically collect metrics
Group by ClusterReport Hours vs. Jobs
AutomationZero touch deploymentCloudFormationAWS Service CatalogAuto ProvisioningJust-in-time setup