RCECloud: Earned Value Analysis

Task	Planned Value	Earned Value	Actual Cost	Cost Variance	Schedule Variance	Estimate at Completion	Budgeted at Completion	Variance at Completion	CPI	SPI	CV %	SV %	TCPI
0) RCECloud
1) Offload unwanted or time-consuming services/responsibilities.
1.1) Decommission old/unused hardware.
1.2) Automate patching procedure
1.3) Migrate CGA hardware out of data center
1.4) Migrate/remove Redmine
1.5) Migrate RT to secondary provider or utilize RT engineers to upgrade our installation. Investigate providers.
1.5.1) Research which provider or hosted solution is best.
1.5.2) Work with provider to upgrade and maintain RT installation.
1.5.3) Make an inventory of all of our ‘RT’ scrips and discard those which are no longer necessary for migration.
1.6) Migrate tape backup archive to BackBlaze.
1.6.1) Get BackBlaze certified with OGC/HUIT
1.6.2) Buy and install host which performs backups
1.6.3) Work with BackBlaze to find the most efficient way of pushing this data
1.6.4) Use HashBackup
1.6.5) Measure performance and reliability of process; backup data to tapes and BackBlaze
1.6.6) Find out what we need to backup, and when
1.6.7) Re-write documentation on backing up user, project data, and restoration to conform to BackBlaze backup system.
1.7) Migrate all web services, except ShinyServer, to Heroku, including DVN.
1.7.1) Migrate and decommission Jenkins, reuse.
1.7.2) Work with Salesforce/Heroku to build an HMDC Heroku Enterprise contract
1.7.3) Create a basic-enterprise HMDC Heroku account with some limits.
1.7.4) Give Raman access to a basic Heroku account and ask him to deploy Dataverse Network to its and its components.
1.7.5) Certify Heroku with OGC
1.7.6) Create a Heroku account and deploy Latanya’s PHP application to it.
1.7.6.1) If successful, work with Latanya to migrate one of her websites entirely to Heroku to test the viability of this environment.
1.7.7) Work with Heroku to determine enterprise pricing for the amount of web sites we host.
1.7.8) Determine the best method to migrate our users permissions to Heroku: How much CPU time should they get? How much memory should they get? Should they be allowed to provision as many websites as they want, etc.
1.7.9) Determine the best method for file storage on Heroku — using an S3 bucket, connecting to our local NetApp, etc.
1.7.10) Work with Gary and Cris to decide how much support we provide for users in the Heroku environment: Do we help them setup GitHub repositories? Do we help them with basic PHP programming fundamentals, permissions?
1.7.11) Migrate all sites to Heroku. Provide a deadline.
1.7.12) Ask DVN to migrate their entire toolchain to the Cloud. Provide a deadline.
2) Work with Microsoft to build RCE-in-Azure launching pad
3) AWS/AZURE/Cloud migration preparations complete
4) Test RCE/Azure PoC with a small user-base
5) Re-design authentication/authorization
5.1) Create a small web app — my.hmdc — to sync accounts.
5.2) Work with Harvard IAM to develop security procedures around LDAP re-writes.
5.3) Develop re-write engine to sync Harvard Key and local LDAP accounts.
5.4) Re-write documentation wrt. account creation and meet with Soner’s team.
6) Authentication re-design completed
7) Re-design RCE/Azure to support multi-cloud submission
7.1) Use docker containers for applications rather than stored images, export with XPRA.
7.2) Install Rancher environment to Clouds
7.3) Install Kubernetes environment to Clouds
7.4) Integrate Harvard Open Nebula
7.5) Build REDIS queue and worker to distribute tasks to multiple clouds
7.6) Build docker images for HTCondor negotiator, collector, and schedd for AWS with redundancy
7.7) Redesign RCE/Azure application to submit jobs to redis queue and manage jobs through the redis queue and associated workers
7.8) Scale NodeJS web service appropriately using npm-cluster/multiple workers
7.9) Implement tests for NodeJS code
7.10) Use Terraform to create abstract cloud configurations for AWS, Azure, Google
7.11) Auto-build and test docker images
7.12) Implement monitoring for RCE Cloud
7.13) Allow mounting of home directories - CI3/CI4 to Azure and CI3/CI4 to Amazon
8) SID public beta release
9) Resolve any outstanding bugs acquired from BETA release
10) Begin to migrate users to RCE Cloud
11) Migrate most services to Cloud
11.1) Slowly de-commission execute nodes, sell or move to recycling.
11.2) De-commission RCE Login nodes, sell or cycle, migrate licenses to Cloud.
11.3) With all web apps migrated to Heroku, turn off and recycle www-* hosts.
11.4) Migrate pertinent fs-4 tasks to another host, turn off and recycle fs-4.
11.5) With RT migrated to cloud provider, turn off RT and recycle.