This document provides an overview of what's new in Documentum 7.3, including:
- The introduction of stateless Documentum and application containers using Docker for continuous delivery and faster upgrades.
- Enhancements to security features like SAML 2.0 single sign-on and updates to the Java Development Kit update tool.
- Expanded REST APIs and support for CMIS 1.1 for improved integration and user enablement.
- Optimization and automation tools aimed at reducing the effort required for patches, upgrades and overall administration.
3. Recent News
• Business as usual until close: 90-120 days
• Roadmaps have always been under NDA and subject to change
• Momentum ECN for FAQs and ongoing communications
• Follow SPARK blog for executive updates - sparkblog.emc.com
“Dell EMC and OpenText are committed to
supporting customers through the transition with
world-class support. In addition, it’s OpenText’s
intention to maintain and invest in ECD’s portfolio to
protect customers’ investments in their current
products and solutions.”
- Mark Barrenechea, CEO & CTO, OpenText
You may have seen this text in other sessions, but it is worth displaying again: this session includes forward looking statements, and anticipated product features, and while EMC makes every attempt for these to be accurate, they are subject to change and should be used for the purposes of discussion.
Also, please note that all of this information is subject to the EMC Non-Disclosure Agreements in place with your organizations.
TOP AREAS OF CUSTOMER CONCERN DRIVING DEVELOPMENT
Effort, Risk, Disruption of Upgrades / Patches
Total Cost of Ownership within SDDC
Trust and Security
New User Engagement
Analytics
Started with what customers were asking for
Optimizations for the current installers
And Best Practrices to prepare and modify repositories
No big changes, no disruption
Platform releases are 18 to 24 months. Where would customers need to be in two years? Could we go beyond current expectations?
RE-INDEXING CHALLENGE
Each Database change requires re-indexing usually blocking database
Sequential scripts adds up the hours quickly
The answer: ELIMINATE FORCED CHANGES
No Changes forced during an upgrade; anything for new features can be post upgrade
No disruption due to big Architectural changes
Enables: STATELESS DOCUMENTUM
Decouple from Repository for Binary-Only updates
Upgrade = Point Content Server 7.3 to D7.2 Repository
Eliminates Downtime due to Reindexing
Upgrade Time is independent of Repository size
CHALLENGES
COMPLEX PLANNING to REDUCE DOWNTIME
ROLLBACK
PRODUCT INSTALLERS
CONFIGURATION
PORTABILITY
UNIFORM ACROSS STACK (UPGRADES and PATCHES)
DEPLOYMENT OF NEW STATELESS DOCUMENTUM
Benefits of CONTAINERS
Virtual Application Container
Production Ready – No Installers, PreConfigured
Quick deployment to any cloud system or OnPrem
System-as-is with various deployment strategy
Quick deployment and rollback
Proper selection of containers is necessary for:
Separation of concerns
Automation
Efficiency
Broad ecosystem
Containers are much more efficient than VMs:
Containers are isolated but share OS and where appropriate bins and Libraries. No OS is necessary
Fewer resources
Less restart time
Many apps and app variations can be supported by the same runtime engine
Uniform upgrade and patch process across all Documentum solution stack
Very easy to understand and implement according to established container paradigms
Developer/POC:
All-in-One Container for Functional Investigations – quickly get to try out the next release – replacing our Developer Edition
Service Images:
Individual Service Containers for Agile Expansion and Load Balancing
Xplore, Rest, DA, etc.
Stateless Documentum
External Content, Database & Isolated Configuration for Fast Patching & Upgrades
We’re also working on
High Availability setups
Migration from On Premise
Container creation via silent installers
Note that we aren’t limiting Docker certification to the Content Server and a few platform components. We will be releasing reference images across the Documentum stack.
You can build out your solution or wait until our Value Office Solution (LS, EPFM, AO, etc) support for Docker post Bedrock.
Note that our references will be based on Linux since Microsoft just certified Docker in WS2016 for Production last week. We’ve been using the beta in our labs, so expect more information on for Documentum and Docker for Windows-SQL Server in the future.
Also note that we will be providing information on how you can develop and a grown at home container solution.
Add Docker containers along side existing Documentum (co-exist)
Deploy extra Docker containers to manage high volumes
Migrate to Docker by turning off original non-Docker instances
1) What are the planned features and improvements in xPlore 1.6 as part of DCTM 7.3 release in November?
I list all features on Bedrock with brief after my signature.
2) Does xPlore 1.6 support Linux RHEL 6.7?
Yes. It has been certified.
3) Does xPlore 1.6 on RHEL 6.7 support Content Server 7.1 on RHEL?
Yes. We had tested xPlore 1.6 and CS 7.1 on Windows during compatiblity testing. We are confidence it works on RHEL as well because the most implementetions are in Java and platform independent.
4) Currently we deployed xPlore 1.5 on Linux RHEL 6.7 with Content Server 7.1 on RHEL. Is there a direct, in-place upgrade path from xPlore 1.5 to xPlore 1.6 on Linux RHEL?
Yes. Pfizer can refer to in-place upgrade steps listed in admin guide.
5) Do we need to reindex the docbases to deploy xPlore 1.6, or xPlore 1.6 can reuse the xPlore 1.5 index? Do we still receive the new features and improvements of xPlore 1.6 if we reuse xPlore 1.5 index?
Reindex the whole docbase is not mandatory. Upgrading from 1.5/1.4 to 1.6 doesn’t require customer to re-index their docbase. The old index can still be used by 1.6. Some features can benefit customer directly, including some query optimization, metadata highlighting (customer has to change their xquery to highlight fields they desire), thesaurus improvements.
But below 2 feature requires user to reindex data if they want to get benefit. They have to turn on them manually via modifying indexserverconfig.xml. By default, those features are turned off. If customer turns on those features, it’s highly recommended customer cleaning their existing data and re-index their docbase to guarantee their data consistency.
• Make the position information in index as optional
• CPS indexes Multilingual content
Thanks,
Dingmeng Xue
• Trust:
o Non-Admin users can run reports
There is typically a need for non-admin users to run and view reports, and system status. The current version of admin interface restricts the xDB admin to have access to admin interface. With this new feature, an access management console is setup that will allow for users with different privileges; access different resources.
o Support user authentication in Admin Console
Follow up above feature. Admin Console can use Documentum content server to do authentication. Admin Console uses white list to store all empowered user but their password would be in Content Server even LDAP. When user login Admin Console, the user credential will be sent to CS. It will allow Admin Console to leverage CS’s SSO capability.
• Improved Content Management
o CPS indexes Multilingual content
In the current offering, CPS identifies the language of the content, file-by-file. So where there are files with multiple language, then such files yield a no match, when the search-key is of a quoted language. Say for instance, if a file was identified as english, and there was french text in that file, then a search for a French key word will not yield a match. With this feature, we can now search on multi-lingual file
English and French would not be a good example, because their tokenization would be very similar. You can mention the content has English and Japanese. Since Japanese has totally different tokenization method. In xPlore 1.5, we will process this document only using one language. Then another language will be ignored and cannot be hit.
o Flexible object routing mechanism:
Offer a new approach that user can store the document to specific collection according to their configuration. Such as user can store the document according to its content type, create date, user name or any other attribute.
o Pause and resume rebuilding index
Indexing of content requires a fair share of the CPU power and it helps to have control as to when to ask for a bigger share of the CPU power and when to let-go of the CPU share; like in giving away processing capability to serve the user needs. This feature - Rebuild index on Admin Console, help users manage rebuild index task on each collection. User can continue previous work to process multipath index instead of redo it. A progress bar in the admin interface, will show the statistics of indexing process.
o Support SKOS format two-way expansion
With a two-way expansion, it is possible to use "animals" to find "creatures" and also use "creatures" to find "animals”
o Re-feed specific data set in doc repository via index agent:
In xPlore 1.5, user can upload a txt file containing the list of object ids to IndexAgent and IndexAgent reads this file and do re-feed (loading the data from content server and send them to xPlore Index Agent) one by one. User also can specify the id list through one DQL. This function can help user easily re-feed those objects and fix their content in xPlore Index Server. There would still be failed object because environment issue and customer would want to retry. The shortage was user can trigger a re-feed process but cannot see the failed objects. Another critical design issue was IndexAgent doesn’t store the status of each re-feed item. If customer restarts IndexAgent, the whole process will start from the beginning. It’s unacceptable to customer if they submit a list with thousands of ids. Since Bedrock, IndexAgent will store the status of the list. Once user restarts the system, IndexAgent resume the process from the breakpoint. End user can see the status and export the list of failed ids during re-feed.
• Performance and scalability:
o Optimize multiple term phrase search:
User sometimes uses multiple terms to do the query such as querying “architecture diagram”. The search terms have diverse frequency in whole document set. Some would be very selective but some would be not. So we optimize the search logic and search selective query with higher priority.
o Improve performance of folder descend query:
Original folder descendent query needs to load all its descendant folders which is very expensive operation if parent folder has many descendant folders. So we optimize the process in this release and use cache to speed up folder navigation.
o Make the position information in index as optional
Position information needs additional storage in full-text index. We offer an option to turn off it. The benefit is index size will reduce to one fifth. Smaller index file, faster search speed. The sacrifice is phrase match query cannot be supported. Such as querying “enterprise content management”, we only return document which contains exactly 3 words together with same order. If no position information, it will return any document as long as it contains those 3 words no matter their position.
o Import and update large volume of thesaurus file
Import large thesaurus file is very heavy option in old xPlore version even making system to be out-of-memory. We optimize file parser in this version and it can handle the big file smoothly.
o Queries derived from cache:
ACL is using cache to speed up calculation. However customer would have no idea to adjust the size of cache. We will log the cache hit ratio. Administrator can check the log to understand whether the performance is impacted by the cache.
o Cut-off notification for wildcard execution
Wildcard search has internal cut-off mechanism to mitigate risk of unselective query. We expose new API on DFC level to tell user whether cut-off applies for this query and how many terms are selected for this query.
• Reduced TCO:
o Configure multiple CPS instances via Admin Console uniformly
This feature provides an easy way to apply and keep all CPS instance configuration keep consistently. Configuration will be pushed to different cps instance after click “save” button. These CPS instances have to be registered on Admin Console.
o Preserve configuration profiles during upgrades
During upgrade, all customer’s configuration in index server, xdb, Index agent, admin console, and CPS will be preserved.
o Certify Docker Deployment
Generate docker image contains xPlore out-of-box configuration. and certify its quality running in Docker container
How many people are using or have used CMIS with Documentum?
Browser Binding, bulkUpdate properties and Append to a Content Stream
Content Management Interoperability Services 1.1 :
Browser Binding
JSON Payload. A binding is the protocol a client uses to talk to a CMIS server. CMIS 1.0 supported two bindings, Web Services (SOAP) and AtomPub (RESTful XML), the latter being the most performant and the most popular. But if you’ve ever looked at the XML that comes back from a CMIS AtomPub call you know how verbose it can be. The Browser Binding is based on JSON, so the payloads that go between client and server are smaller, making it the fastest of the three bindings. The original purpose of the Browser Binding was to make it easy for those building “single page” web apps or doing other work with CMIS via client-side JavaScript, but I think apps of all types will move to the Browser Binding as quickly as possible simply because it is easier to work with
Bulk Updates
CMIS 1.1 adds a new feature that makes mass changes more performant. Instead of iterating over a list of objects, changing and saving each one, you can define a set of property changes and make those against an entire collection, which is much more efficient.
Append to Content Stream
A challenge with any ECM project is how to move large files into the repository. The new append to content stream feature in CMIS 1.1 allows you to send files to the repository in chunks which could be a key to addressing that challenge.
REST services are fundamental to most new client development. How many people have used or are using our REST Services today?
Are there any Services you have found fundamentally missing for your needs?
Three main areas of focus for new features have gone into the Bedrock release: Advanced Search, User Management & Access, Advanced type support.
Our plan is increase the frequency of REST service releases (disengage from the Platform releases) to meet internal and external customer demands after Bedrock. We will continue to align major releases with the platform, however more agile incremental services will be made available.
Some of the areas we are looking at include Workflow Services, RPS, Content Transformation Services and collaboration.
The Vision for REST is to not only provide what is fundamentally provided by DFS today, but to exceed DFS capabilities.
While EMC is committed to REST development; DFC,DFS,CMIS is not going away
I’m certain most of you are aware of the major upheaval in the browser world.
Java:
JRE Security Fixes force continuous pressure on EMC Certification and Customer IT teams
Browsers:
NPAPI removed from top Web Browsers eliminates application to UCF bridge
HTML5:
Modern HTML language prohibits advanced interaction with user file system
over NPAPI being removed and we are all too painfully aware of what this means for clients leveraging advanced content transfer capabilities.
The fundamental problem is security, giving a Web Browser capability to read and write files on the client machine.
The solution today is either to force the user to save files/select them for checkin or involve browser extensions/helper objects and native client that must be installed/trusted on the client machine.
Our solution: UCF stays and depending on the client will be UCF Java or .NET, but the clients will be modified to leverage the new Java Applet Free download manager…which means …no NPAPI browser requirements.
For UCF java there is still the need for client side java, however it is an embedded JRE. The UCF itself did not have issues with the version of Java installed on the users machine, it was actually the applet technology to download the UCF that had all the issues reportedly.
Best practices to direct maintenance releases
Sustain Certifications and Security
Backwards compatibility & fast low risk upgrades
Webtop to D2, xCP functional gaps (unexposed Platform features like federated search)
Support for Oauth and external Trust providers
Windows Docker & Complete Stack Dockerization
Auto-provisioning and scaling
Health check console (syslog collection for support, etc).
Stability: Service Packs
REST & CMIS 1.1 (full)
China and Russia (encryption rules)
Tighter Cloud Services feature support
Native Support for Cloud laas (AWS, Azure, CloudFoundry, Compute)
Industry/Government Standards Compliance (508, ITAR, FEDRAMP, Common Criteria, FSTEC, IL3, ISO, PCI, PrivacyShield, SEC)
XML Apps (JSON)