UC3 M e r r i t t
Curation Micro-Services
“It’s a Series of Tubes”
UC3 M e r r i t t
The Unix philosophy
“Make each program do one thing well”
“To do a new job, build afresh rather than
complicate old programs by adding new features”
“Expect the output of every program to become the
input to another, as yet unknown, program”
“Design and build software … to be tried early”
“Don't hesitate to throw away the clumsy parts and
rebuild them”
— D. L. McIlroy et al., “Unix time-sharing system forward,”
Bell System Technical Journal 57:6, part 2 (1978): 1902
UC3 M e r r i t t
The micro-services “philosophy”
CC
http://www.flickr.com/photos/oskay/265899811/
http://www.flickr.com/photos/elsie/8229790/
UC3 M e r r i t t
Curation micro-services
Metaphors
Assumptions
Principles
Preferences
Practices
Pipeline
Safety through
redundancy
Modularity
The small and simple over
the large and complex
Focus on outcomes, not
means
Lego bricks
Meaning through
context
Granularity
The minimally sufficient
over the feature laden
Complexity through
composition, not addition
Utility through
service
Orthogonality
The configurable over the
prescribed
Policy neutral, platform and
protocol independent
Value through
use (and reuse)
Emergence
The proven over the
(merely) novel
Approach sufficiency through
incrementally necessary steps
Stewardship is a
relay
Evolution
Early prototyping, frequent
refactoring
Parsimony
Code to interfaces
UC3 M e r r i t t
Curation micro-services
Mode
Focus
Value
Accretion
Value
Utility
Annotation
Visibility
Notification
Accessibility
Access
Derivation
Transformation
Selectivity
Search
Actionability
Index
Stewardship
Ingest
Epistemology
Characterization
Context
Ontology
Inventory
Reliability
Replication
Fixity
Fixity
Preservation
State
Valence
Stability
Storage
Identity
Identity
Visibility
Interoperation
UI / Access control / Message queuing
Curation
Service
User-facing
Application
Interpretation
Providerfacing
Protection
UC3 M e r r i t t
Design goals
Principle of least surprise
Multiple interface modalities
– RESTful HTTP
– Command line
– Procedural (Java, Perl, Ruby, …)
Linked data
Stable URL references
State
Storage
or
Storage
Object
Version
File
content
node
service
http://example-store/ state/default/1234/3/xyz
The file system is the database
UC3 M e r r i t t
“You say micro, I say macro…”
Service
Tool
Access
Access
ANVL
ARK
bagit.pl
BagIt
CAN
checkm.pl
Checkm
Dflat
ERC
EZID
GhOST
Ingest
Inventory
LockIt
N2T
namaste.pl
Namaste
Noid
Pairtree
ReDD
RUU
Storage
EZID
GhOST
Ingest
Inventory
N2T
RUU
Storage
Convention
ANVL
ARK
BagIt
Checkm
Dflat
ERC
LockIt
Namaste
Pairtree
ReDD
UC3 M e r r i t t
Development roadmap
First wave
Second wave 
Third wave
Fourth wave 
Fifth wave
Sixth wave 
Identity
Inventory
Index
Search
Notification
Annotation
Storage
Ingest / Access
Fixity
Replication
Characterization
Transformation
IDm / Authn / Authz
Metadata standards
Object / collection modeling
Semantic interoperability
Policy / business model development
UC3 M e r r i t t
Ingest process flow
Create identifier
Identity
Identifier
Submitting
user agent
Submit
Ingest
Node
Add version
Notification
Version metadata
Get version
Add version
metadata
Storage
Notification
Node
Version metadata
Get version metadata
Inventory
Version metadata
Node
UC3 M e r r i t t
Ingest implementation
Submitting
user agent
Ingest notification
HTML
form
Batch or
single object
Submission notification
Job
metadata
Submitter
Queue
Servlet
Implicitly multi-threaded
Job
payload
Zookeeper dæmon
Consumer
Dæmon
Explicitly multi-threaded
Ingester
Servlet
Implicitly multi-threaded
Storage
UC3 M e r r i t t
Questions?
silverpipes.jog / firstpresmacomb.org
UC3 M e r r i t t
More information
UC Curation Center (UC3)
http://www.cdlib.org/uc3
Micro-service specifications
https://confluence.ucop.edu/display/Curation
Digital curation group
http://groups.google.com/group/digital-curation
UC3
Stephen Abrams Erik Hetzner
Patricia Cruse
Greg Janée
Scott Fisher
John Kunze
Margaret Low
Mark Reyes
Perry Willett
David Loy
Tracy Seneca
Isaac Rabinovitch Marisa Strong