An overview of enhancements and additions to high availability and site resilience capabilities since the Exchange 2010 release.

Reduction in IOPS
Reduction in IOPS enables us to leverage larger disks in terms of capacity and IOPS as efficiently as possible. 

In the previous version of exchange, a low checkpoint depth was kept for fast failover, due to which IOPS of the passive copy was equal to the Active copy.

Exchange 2016 Allow fast failover allowing us to keep the deep checkpoint depth of 100 MB as a result IOPS of Passive copy has come down to 50% compare to active copy.
Managed availability
Managed Availability integrate active monitoring and the Exchange 2016 high availability platform.

With Managed Availability system can make a determination on when to fail over a database based on service health.
  • Probe Engine: is responsible for taking measurements on the server and collecting data
  • The monitor : the monitor looks for the various different patterns on all the collected measurements, and then it decides whether something is considered healthy
  • Responder engine : responsible for recovery actions
Managed availability is implemented in the form of two services:
  • Exchange Health Manager Service (MSExchangeHMHost.exe) 
  • Exchange Health Manager Worker process (MSExchangeHMWorker.exe)
Managed Store

The Managed Store is the name of the newly rewritten Information Store processes in Exchange 2016. The new Managed Store is written in C# and tightly integrated with the Microsoft Exchange Replication service (MSExchangeRepl.exe) to provide higher availability through improved resiliency.
Support for multiple databases per disk Exchange 2016 includes enhancements that enable you to support multiple databases (mixtures of active and passive copies) on the same disk, thereby leveraging larger disks in terms of capacity and IOPS as efficiently as possible.
AutoReseed Automatic reseeding capability enables you to quickly restore database redundancy after disk failure. If a disk fails, the database copy stored on that disk is copied from the active database copy to a spare disk on the same server. If multiple database copies were stored on the failed disk, they can all be automatically reseeded on a spare disk. This enables faster reseeds, as the active databases are likely to be on multiple servers and the data is copied in parallel.
Automatic recovery from storage failures This feature continues the innovation introduced in Exchange 2010 to allow the system to recover from failures that affect resiliency or redundancy. Exchange 2016 now includes additional recovery behaviors for long I/O times, excessive memory consumption by MSExchangeRepl.exe, and severe cases where the system is in such a bad state that threads can't be scheduled.
Lagged copy enhancements Lagged copies can now care for themselves to a certain extent using automatic log play down. Lagged copies will automatically play down log files in a variety of situations, such as page patching and low disk space scenarios. If the system detects that page patching is required for a lagged copy, the logs will be automatically replayed into the lagged copy to perform page patching. Lagged copies will also invoke this auto replay feature when a low disk space threshold has been reached, and when the lagged copy has been detected as the only available copy for a specific period of time. In addition, lagged copies can leverage Safety Net, making recovery or activation much easier.
Single copy alert enhancements The single copy alert first introduced in Exchange 2010 is no longer a separate scheduled script. It's now integrated into the managed availability components within the system and is a native function within Exchange.
DAG network auto-configuration DAG networks can be automatically configured by the system based on configuration settings. In addition to manual configuration options, DAGs can also distinguish between MAPI and replication networks and configure DAG networks automatically.