3.1 SERVER ROLES In SharePoint 2013 servers were role agnostic – that is, any server could provide any service by stopping or starting services, and the documented guidelines were pretty much around four to five general roles (caching, request management, web servers, batch processing, and specialized workloads). In SharePoint 2016, roles are defined in code instead of in documentation, and there are basically three roles: - User Services - any request initiated by an end user is processed by a User Service machine. In 2013, a user request could traverse the entire topology, in 2016, the new role model is designed to simplify the performance and latency issues around that use case and simplify this scenario. - Robot Services – any request not initiated by an end-user - timer jobs, indexing, etc. - Caching Services – basically provides support for caching. 3 Enter the definition of MinRoles. MinRoles allow you to define the role of the server as you’re laying down the bits (either through the configuration wizard, PSConfig, or PowerShell), and apply the appropriate SharePoint logic based on the role. Here are the new MinRoles: - WebFrontEnd (corresponds to UserServices above) - Application (corresponds to Robot Services above) - DistributedCache (corresponds to Caching Services above) - SpecializedLoad (this setting allows you to break the model and lay down pretty much any configuration of services you like – just like you did with SharePoint 2013) In the configuration wizard there are a couple of additional choices: - Search – allows you to create a Search Server - Single Server – Allows you to lay down everything on a single server as you did with SharePoint 2013. It’s not clear if the naming of these will remain the same or if all of these will make it to the final product but the MinRole model is clear and the topology has clearly changed for the better to minimize the traversal of a user request across multiple servers. 3.2 HEALTH ANALYZER Based on the above Server Roles, the Health Analyzer will detect any deviation from the server role norm and allow you to fix the deviation (with the exception of SpecializedLoad). For example, if you have a WebFrontEnd that had an additional service installed that is outside the definition, the Health Analyzer will detect the deviation, alert you and allow you to “fix” the deviation by removing the service. 3.3 PATCHING This topic was one of the biggest changes, and it is all due to learnings from delivering on SharePoint Online SLA’s. Those SLA’s drove the requirement to deliver “in-place, on-line, middle-of-the-day, transparent to the user, no need to stop and start services” to apply patches. In the past, SharePoint QFE's consisted of a number of MSI's and MSP's. 37 MSI’s/MSP’s plus an additional 18 for each language pack. For 2016, the product team has now minimized this down to 2 MSI's/MSP's for a total of 4 plus 1 for each language pack. This provides a much smaller patch and a smaller footprint, the updates are smaller and now execute with zero downtime. The reduced patches also reduces the number of testing scenarios – bottom line – patching can now be done during the day, and your users won’t notice. 3.4 DISTRIBUTED CACHE There are also some significant improvements to the way the Distributed Cache works by reducing NTLM authentication overhead by switching off NTLM authentication such that the cache doesn’t have to authenticate on every request. This significantly improves performance of the cache. 4 3.5 NEW BOUNDARIES AND LIMITS These new limits are just guidelines today – goals that the dev team hopes to achieve. We sincerely hope they are able to exceed these boundaries. Content Database Size Site Collections Per Content Database List Threshold Max File Size Indexed Items Content database sizing into TB’s 100,000 site collections per content database Increased List Threshold >5000 MaxFile Size increases to 10GB and removed character restrictions 2x increase in Search scale to 500 million items 3.6 FILE PERFORMANCE For File upload, they are removing the SOAP over HTTP protocol and replacing that with the BITS (Background Intelligent Transfer Service) protocol, which should significantly improve the upload performance. 3.7 FAST SITE CREATION In typical Site Creation, a site is created, and features are activated one by one, which is a very time consuming process, Fast Site Creation uses a site template with everything already pre-provisioned and then does a site copy of the template at the content database level. This is a significantly faster process, and one that is already being used in SharePoint Online today. 3.8 TRAFFIC MANAGEMENT Traffic Management is a new endpoint running on the web servers that establishes an affinity between the web servers and any load balancers (could be Big IP, F5, etc.). It effectively acts as an intelligent transport layer and routes the traffic to the most efficient server based on usage, etc. This feature may or may not be delivered on premise, but is currently being offered today with SharePoint Online. 3.9 USER PROFILE SERVICE The User Profile Sync is no longer provided based on an internal FIM service, but is now provided bidirectionally using an external FIM service, or unidirectional using native AD synchronization. 3.10 PROJECT SERVER The Project Server Databases now become first class SharePoint citizen in that they are now included as part of the SharePoint content Database. This simplifies the management of databases, but you still have to license this the old fashioned way – separately from SharePoint. 5 3.11 DURABLE LINKS Saving one of the best for last – the URL for a document is now based on a resource ID, not the name, so not matter what you name the document or where you move it, the URL always works. For example if the document was named “Sample Document –Draft.docx” and later named “Sample Document”, any URL pointing to it would still be able to find it – very cool stuff. 4 SUMMARY This is just a first sneak peek into the new features of SharePoint 2016 and really just focused on the IT Professional features of the on premise product. Clearly Microsoft has learned a lot around delivering a world class service and around delivering on very high SLAs around that service. Much of those learnings have made their way into the product to all of our advantage.


Popular Posts