- Async/Workflow Updates
- Async servers will now monitor available jobs and pull extra workloads based on available server resources.
- Workflow processing throughout has been increased from 25% to 50%
- Action Updates
- In CRM Vega actions were primarily only useful to developers because they were only accessible via code. As of Carina actions will be callable from dialogs and workflows. This allows actions to be used more as a library of API of business functions
- Actions are akin to stored procedures allowing for business logic to be exposed to clients instead of clients having to each implement their own version of business logic
- This provides a more consolidate programming model and maintains the application domains boundaries
- If business logic is spread out between two or more integrated applications it becomes a maintenance overhead and creates opportunities for errors and issues
- Currently only basic data types are supported as parameters to actions
- Future releases will be able to use complex data types
- Updated touch friendly navigation menu
- If you had a system with a fair amount of menu options you probably cursed the sky if you ever accidently moved your mouse off the menu as it would disappear. Now you can navigate in pure bliss with the updated menu UI.
- New OData v4 preview end point available
- Turned off by default
- Should not be used for production
- Why should you care?
- OData end point has much less communication overhead than the existing SOAP end point. Can provide a performance enhancement in high volume data situations.
- Possibly available in CRM 2015 Fall Release
- SDK will be fully implemented via OData end point
- One of the great benefits of this is that the OData protocol has much less overhead than SOAP and can help improve integration performance.
- Data Services
- Multi-entity search was introduced to the web client in CRM Vega
- Multi-entity search simply allows a user to search multiple entities without having to first navigate to each entity before performing a search
- Quick Find by default uses a Starts With query which does not provide expects results
- Wild cards can be used but performance a full table scan in the database because no indexes can be used for wild card searches
- As of CRM Vega UR1 full-text search will be available for on-premises customers
- It will be turned off by default until enabled by an administrator
- Quick find will operate as normal without full-text search turned on
- Wild card search performance significantly increased on multiple word searches
- Wild card characters are no longer required and stripped off if present
- Quoted strings perform an exact match search
- Attachments are not indexed or searchable by full-text search
- As of today the quick find and multi-entity search are internal API’s and do not have any support for search operators
- For customers wanting to customize the search results the recommended approach is to call the quick find view and pass the desired parameters to it via the SDK
- In the future a public search API will be provided to allow customers to customize the search function
- Multi-entity search was introduced to the web client in CRM Vega
- Data and Time (You can now get a date with CRM)
- Date and Time as implemented today was originally meant to represent user local time for things that happened at a single point in time. It worked well for scenarios such as created or modified dates but fell short when it came to things such as birthdays.
- New Behaviors
- The CRM 2015 Spring release will implement the Date Only field type which will always store time as 12:00 AM providing a resolute date field that doesn’t convert based on the user’s timezone
- CRM will now support date values going back to 1/1/1753
- Excel still only supports dates going back to the year 1900 and will convert any date earlier to a string
- In the future the product team is looking at timezone support
- CRM Online Service
- Globally available in over 131 markets and 44 languages
- Service deployments twice a year
- Ongoing Infrastructure enhancement behind the scenes
- Introducing the Microsoft EDGE network
- EDGE is a global private Microsoft network that has a 20 ms access time from any global hub
- Optimization of the “middle mile”
- In the past there has always been a network latency issue for CRM online
- When a user accesses the service they must first go through their ISP to for domain resolution
- Next the ISP would find a route to get to the CRM services which could potentially be a few hops before reaching the Microsoft data center
- Now when a user attempt to access the service they will not have to make the hops that were once required which will provide increased performance
- Estimated performance improvement is around 25% for customers in more distant regions
- The EDGE network is not confined to Microsoft data centers
- Already have about 400 locations called “points of presence” near major business hubs
- Customers don’t have to do anything to benefit from the EDGE service as it has been deployed behind the scenes for all customers
- The goal is to have all CRM online customers worldwide will be on the EDGE network by July/August 2015
- Internal Application Routing
- Previously CRM online relied on external ISP DNS servers to keep up with changes to what servers CRM was running on in the Microsoft data centers
- This proved problematic because while Microsoft would publish a policy for ISP DNS servers to refresh their cache every 30 seconds some of them would ignore the policy and only refresh every 24 hours
- This was perceived by customers as an outage, but the problem was that they never received the updated DNS for the new server location
- The CRM online team corrected this by introducing Property Front Door servers which act as an internal DNS for CRM services
- Now when a CRM service is moved to another server the PFF server will reroute users to the proper server without having to publish a new DNS record. This removes the dependency on ISP’s refreshing their DNS in a timely manner
- The Property Front Door servers refresh their cache every 10 seconds
- Minimal Downtime Updates
- A CRM online scale group is composed of 3 components
- F5 Load Balancers
- Application Servers
- Database Servers
- This architecture is easy to scale out when resources are low
- Scale groups can be added and customers moved between them
- The problem is that when a customer moves between scale groups
- The CRM service is stopped
- The database is detached
- The database is moved and imported into the new scale group
- The service is restarted
- During this period the customer is down
- The online team is moving to a new architecture to prevent down time when moving scale groups or performing upgrades
- The new methodology revolves around decoupling the application servers from the database servers and removing the requirement for each version of CRM running in a separate database
- With the new methodology a single database server can host multiple versions of CRM and org database can be updated independently of each other
- This allows for in place upgrades without the need of moving database servers from one scale group to another
- Also by decoupling the applications servers this can scale out better than before
- For large enterprises that require higher performance CRM online team can define infinity groups which can dedicate a set of application and database servers to a particular customer
- A CRM online scale group is composed of 3 components
- Improvements for CRM Online Spring 2015 Preview
- Sandbox Support for Preview Instances
- Sandbox Support allows for the following
- Reset
- Delete an org and start back from a fresh org
- Copy
- Copy one instance and create another
- Admin Mode
- Locks everyone except administrators out of the system
- To be used while administrative tasks are being performed
- Reset
- Preview Storage
- Preview instances don’t consume paid storage space.
- Preview instance storage is free but not infinite
- Customers given about a 25% bigger over what was copied from their production organization
- Preview instances can be updated on demand
- While production instances must be scheduled for an update admins will be able to update preview instances on demand
- Everything discussed above will be available in the Carina deployment
- Access to sandbox features and functionality via the SDK is on the roadmap but is a little further out.
