Windows Server 2016 | EPFL Graph Search

Windows Server 2016 is the eighth release of the Windows Server operating system developed by Microsoft as part of the Windows NT family of operating systems. It was developed alongside Windows 10 and is the successor to the Windows 8.1-based Windows Server 2012 R2. The first early preview version (Technical Preview) became available on October 1, 2014 together with the first technical preview of System Center. Windows Server 2016 was released on September 26, 2016 at Microsoft's Ignite conference and broadly released for retail sale on October 12, 2016. It was succeeded by Windows Server 2019 and the Windows Server Semi-Annual Channel. Windows Server 2016 has a variety of new features, including Active Directory Federation Services: It is possible to configure AD FS to authenticate users stored in non-AD directories, such as X.500 compliant Lightweight Directory Access Protocol (LDAP) directories and SQL databases. Windows Defender: Windows Server Antimalware is installed and enabled by default without the GUI, which is an installable Windows feature. Remote Desktop Services: Support for OpenGL 4.4 and OpenCL 1.1, performance and stability improvements; MultiPoint Services role (see Windows MultiPoint Server) Storage Services: Central Storage QoS Policies; Storage Replicas (storage-agnostic, block-level, volume-based, synchronous and asynchronous replication using SMB3 between servers for disaster recovery). Storage Replica replicates blocks instead of files; files can be in use. It's not multi-master, not one-to-many and not transitive. It periodically replicates snapshots, and the replication direction can be changed. Failover Clustering: Cluster operating system rolling upgrade, Storage Replicas Web Application Proxy: Preauthentication for HTTP Basic application publishing, wildcard domain publishing of applications, HTTP to HTTPS redirection, Propagation of client IP address to backend applications IIS 10: Support for HTTP/2 Windows PowerShell 5.

Robust Collaborative Learning with Linear Gradient Overhead

Rachid Guerraoui, Nirupam Gupta, John Stephan, Sadegh Farhadkhani, Le Nguyen Hoang, Rafaël Benjamin Pinot

Collaborative learning algorithms, such as distributed SGD (or D-SGD), are prone to faulty machines that may deviate from their prescribed algorithm because of software or hardware bugs, poisoned data or malicious behaviors. While many solutions have been ...

PLMR2023

Practical Byzantine-resilient Stochastic Gradient Descent

Sébastien Louis Alexandre Rouault

Algorithms are everywhere.The recipe for the frangipane cake is an algorithm.If all the listed ingredients are available and the cook is sufficiently deft, after a finite number of small, well-defined steps a delicious dessert will exit the oven.Now, what ...

EPFL2022