Transcript 2 - Center

Cisco - MS
Javier Echave
Michael Faden
1
Transforms
the datacenter
Empowers
people-centric IT
Unlocks insights
on any data
Enables
modern apps
CONSISTENT
PLATFORM
Windows Server
Windows Azure
Major datacenter
CDN node
Live sub-region
Announced sub-region
Partner-operated sub-region
New levels of
performance and
cross-platform
support
IT demands
Support for demanding,
critical business workloads
Increased flexibility
to move resources
and workloads
Leverage existing crossplatform investments
Windows Server 2012 R2 delivers
Enterprise-class scale
and performance.
Virtual machine mobility.
First-class citizen guest
support for Linux.
Massive scalability for the
most demanding workloads
Hosts
• Support for up to 320 logical processors
& 4TB physical memory per host
• Support for up to 1,024 virtual machines
per host
Clusters
• Support for up to 64 physical nodes &
8,000 virtual machines per cluster
Virtual Machines
• Support for up to 64 virtual processors and
1TB memory per VM
In Guest Non-Uniform
Memory Access
• Projects NUMA topology onto a virtual
machine
• Allows guest operating systems and
applications to make intelligent NUMA
decisions
• Aligns guest NUMA nodes with host
resources
• Workloads such as SQL Server 2012, or IIS
8.0 can take advantage of Guest NUMA
High levels of performance for
key Microsoft workloads
SQL Server 2012
• 64 vCPU support drove 6x performance
increase over previous version of Hyper-V
• 6.3% overhead compared with physical
Exchange 2013
• Virtualized 48,000 simulated users on a
single Hyper-V host across 12 VMs, with
low response times
SharePoint 2013
• Scaled to over 2 million heavy users at 1%
concurrency, across 5 VMs on a single
Hyper-V host
SharePoint
Workload
Hyper-V Virtual
CPUScalability
Scalabilityon
Windows
Server
with Hyper-V
with
OLTP2012
Workloads
VHDX & Support for
Advanced Format Drives
Capabilities
• Improved performance of virtual hard disks
on 512e disks
• Ability to host virtual hard disks on native
4 KB disks
Benefits
• Reduces impact of 512e disks on virtual
hard disk stack
• Workloads complete more quickly
VHDX Provides Increased
Scale, Protection & Alignment
Features
• Storage capacity up to 64 TBs
• Corruption protection during power failures
Block Allocation
Table (BAT)
User data blocks
Sector bitmap blocks
• Optimal structure alignment for large-sector
disks
Benefits
• Increases storage capacity
• Protects data
• Helps to ensure quality performance on
large-sector disks
User metadata
Metadata table
File metadata
Online VHDX Resize provides
VM storage flexibility
Expand Virtual SCSI Disks
1. Grow VHD & VHDX files whilst attached
to a running virtual machine
2. Then expand volume within the guest
Shrink Virtual SCSI Disks
1. Reduce volume size inside the guest
2. Shrink the size of the VHD
or VHDX file whilst the VM is running
Token-based data transfer
within the storage array
Benefits
• Rapid virtual machine provisioning and
migration
• Faster transfers on large files
• Minimized latency
• Maximized array throughput
• Less CPU and network use
• Performance not limited by network
throughput or server use
• Improved datacenter capacity and scale
Access Fibre Channel SAN
data from a virtual machine
• Unmediated access to a storage area
network (SAN)
• Hardware-based I/O path to virtual hard
disk stack
• N_Port ID Virtualization (NPIV) support
• Single Hyper-V host connected to different
SANs
• Up to four Virtual Fibre Channel adapters
on a virtual machine
• Multipath I/O (MPIO) functionality
• Supports Live migration
Provides Near-Line Rate to a
VM on Existing Hardware
vNIC
• vRSS makes it possible to virtualize
traditionally network intensive physical
workloads
• Extends the RSS functionality built into
Windows Server 2012
Node 2
Node 0
Node 1
0
1
2
3
0
1
2
3
Node 3
• Maximizes resource utilization by spreading
VM traffic across multiple virtual processors
• Helps virtualized systems reach higher
speeds with 40 Gbps and 100 Gbps NICs
• Requires no hardware upgrade and works
with any NICs that support RSS
Incoming
packets
Increased efficiency of network
processing on Hyper-V hosts
Without VMQ
• Hyper-V Virtual Switch is responsible for
routing & sorting packets for VMs
• This leads to increased CPU processing, all
focused on CPU0
With VMQ
• Physical NIC creates virtual network queues
for each VM to reduce host CPU
With Dynamic VMQ
• Processor cores dynamically allocated for a
better spread of network traffic processing
Integrated with NIC hardware
for increased performance
• Standard that allows PCI Express devices to
be shared by multiple VMs
Virtual Machine
VM Network Stack
Synthetic NIC
• More direct hardware path for I/O
• Reduces network latency, CPU utilization for
processing traffic and increases throughput
• SR-IOV capable physical NICs contain
virtual functions that are securely
mapped to VM
• This bypasses the Hyper-V Extensible
Switch
• Full support for Live Migration
Hyper-V
Extensible Switch
Virtual Function
Achieve higher levels of
density for your Hyper-V hosts
Windows Server 2008 R2 SP1
• Introduced Dynamic Memory to enable
reallocation of memory automatically
between running virtual machines
Maximum
memory
Maximum
memory
Minimum
memory
Enhanced in Windows Server 2012 & R2
• Minimum & Startup Memory
• Smart Paging
• Memory Ballooning
• Runtime Configuration
Hyper-V
Physical
memory
pool
Utilize disk as additional,
temporary memory
Maximum
memory
Maximum
memory
Hyper-V Smart Paging
• Reliable way to keep a VM running when
no physical memory is available
Minimum
memory
Minimum
memory
• Performance will be degraded as disk is
much slower than memory
Used in the following situations:
Hyper-V
• VM restart
• No physical memory is available
• No memory can be reclaimed from other
virtual machines on that host
Physical
memory
pool
Maximum
memory
Minimum
memory
Features
• Uses resource pools
• Compatible with all Hyper-V
operations
• Unaffected by virtual machine
movement
• Uses Network Metering Port
ACLs
Benefits of Resource
Metering
• Easier to track virtual machine
use
• Can be used to aggregate data
for multiple virtual machines
• Can be used to build accurate
lookback and chargeback
solutions
• Easier to obtain resource use
data
0
Achieve desired levels of
networking performance
Bandwidth Management
Normal
priority
High
priority
Bronze
tenant
Critical
Silver
tenant
Gold
tenant
• Establishes a bandwidth floor
• Assigns specified bandwidth for each type
of traffic
• Helps to ensure fair sharing during
congestion
• Can exceed quota with no congestion
Hyper-V Extensible Switch
Hyper-V Extensible Switch
Gold
tenant
Gold
tenant
Gold
tenant
2 Mechanisms
• Enhanced packet scheduler (software)
• Network adapter with DCB support
(hardware)
Hyper-V Extensible Switch
NIC Teaming
Control allocation of Storage
IOPS between VM Disks
• Allows an administrator to specify a
maximum IOPS cap
• Takes into account incoming &
outgoing IOPS
• Configurable on a VHDX by VHDX
basis for granular control whilst VM is
running
• Prevents VMs from consuming all
of the available I/O bandwidth to
the underlying physical resource
• Supports Dynamic, Fixed
& Differencing
Layer-2 Network Switch for
Virtual Machine Connectivity
Granular In-box Capabilities
• ARP/ND Poisoning (spoofing)
protection
• DHCP Guard protection
• Virtual Port ACLs
• Trunk Mode to VMs
• Network Traffic Monitoring
• Isolated (Private) VLAN (PVLANs)
• PowerShell & WMI Interfaces for
extensibility
Build Extensions for Capturing,
Filtering & Forwarding
Many Key Features
• Extension monitoring & uniqueness
• Extensions that learn VM life cycle
• Extensions that can veto state changes
• Multiple extensions on same switch
Several Partner Solutions Available
• Cisco – Nexus 1000V & UCS-VMFEX
• NEC – ProgrammableFlow PF1000
• 5nine – Security Manager
• InMon - SFlow
Comprehensive feature
support for virtualized Linux
Significant Improvements in
Interoperability
• Multiple supported Linux distributions
and versions on Hyper-V.
• Includes Red Hat, SUSE, OpenSUSE, CentOS,
and Ubuntu
Comprehensive Feature Support
• 64 vCPU SMP
• Virtual SCSI, Hot-Add & Online Resize
• Full Dynamic Memory Support
• Live Backup
• Deeper Integration Services Support
Duplication of a Virtual
Machine whilst Running
Export a clone of a running VM
• Point-time image of running VM
exported to an alternate location
• Useful for troubleshooting VM
without downtime for primary VM
Export from an existing checkpoint
• Export a full cloned virtual machine
from a point-in-time, existing checkpoint of
a virtual machine
• Checkpoints automatically merged into
single virtual disk
VM1 VM2
• Faster live migrations, taking full advantage
of available network
• Simultaneous Live Migrations
• Supports flexible storage choices – iSCSI,
Fibre Channel or SMB for VM’s files
• Requires Failover Clustering if using
iSCSI/Fibre Channel Storage
• No Failover Clustering required if virtual
machine resides on SMB 3.0 File Share
• Can be triggered via PowerShell
Modified
Memory
Storage
Live migration
pages
handle
transferred
moved
setup
VM
MEMORY
Faster, Simultaneous Migration
of VMs Without Downtime
VM
•
Utilizes available CPU resources on the
host to perform compression
•
Compressed memory sent across the
network faster and decompressed on
target host
•
Operates on networks with less than 10
gigabit bandwidth available
•
Enables a 2X improvement in Live
Migration performance
•
Enabled by default but will only operate if
there is spare CPU available to compress
the VM memory.
Modified
Memory pages
Storage
Livecompressed,
migration
handle moved
setup
then transferred
MEMORY
Intelligently Accelerates Live
Migration Transfer Speed
Increased Flexibility through
Live Migration of VM Storage
• Move virtual hard disks attached
to a running virtual machine
• Manage storage in a cloud environment
with greater flexibility and control
• Move storage with no downtime
• Update physical storage available to a
virtual machine (such as SMB-based
storage)
• Windows PowerShell cmdlets
Disk
Disk
Reads
writes
contents
are
andmirrored;
writes
are copied
gooutstanding
to to
new
new
Reads and writes go to the source VHD
changes
destination
are replicated
VHD
Complete Flexibility for Virtual
Machine Migrations
• Increase flexibility of virtual machine
placement & increased administrator
efficiency
• Simultaneously live migrate VM & virtual
disks between hosts
• Nothing shared but an ethernet cable
• No clustering or shared storage
requirements
• Reduce downtime for migrations across
cluster boundaries
Simplified upgrade process
from 2012 to 2012 R2
•
Upgrade from Windows Server 2012
Hyper-V to Windows Server 2012 R2
Hyper-V with no VM downtime
•
Supports Shared Nothing Live Migration
for migration when changing storage
locations
•
If using SMB share, migration transfers
only the VM running state for faster
completion
•
Automated with PowerShell
•
One-way Migration Only
Hyper-V Cluster Upgrade without Downtime
2012 Cluster Nodes
2012 R2 Cluster Nodes
Network Isolation & Flexibility
without VLAN Complexity
• Secure Isolation for traffic segregation,
without VLANs
• VM migration flexibility
• Seamless Integration
Key Concepts
• Provider Address – Unique IP addresses
routable on physical network
• VM Networks – Boundary of isolation
between different sets of VMs
Network/VSID Provider Address Customer Address
• Customer Address – VM Guest OS IP
addresses within the VM Networks
• Policy Table – maintains relationship
between different addresses & networks
Red (6001)
Red (6001)
Red (6001)
192.168.2.13
192.168.2.14
192.168.2.12
10.10.10.10
10.10.10.11
10.10.10.12
Network Isolation & Flexibility
without VLAN Complexity
•
Network Virtualization using Generic
Route Encapsulation uses
encapsulation & tunneling
•
Standard proposed by Microsoft, Intel,
Arista Networks, HP, Dell & Emulex
•
VM traffic within the same VSID routable
over different physical subnets
•
VM’s packet encapsulated for
transmission over physical network
•
Network Virtualization is part of the
Hyper-V Extensible Switch
Network Virtualization Packet Flow
Blue1 sending to Blue2
Hyper-V Switch
Hyper-V Switch
Network Virtualization
Network Virtualization
ARP TABLE
10.10.10.11
34:29:af:c7:d9:12
Network Virtualization Packet Flow
Blue1 sending to Blue2
VSID
VSID
Hyper-V Switch
VSID ACL Enforcement
VSID
IP Virtualization
Policy Enforcement
Routing
VSID
IP Virtualization
Policy Enforcement
Routing
Bridge Between VM Networks
& Physical Networks
• Multi-tenant VPN gateway built-in to
Windows Server 2012 R2
• Integral multitenant edge gateway for
seamless connectivity
• Guest clustering for high availability
• BGP for dynamic routes update
• Encapsulates & De-encapsulates
NVGRE packets
• Multitenant aware NAT for
Internet access
Integrated Solution for
Network Card Resiliency
•
Vendor agnostic and shipped inbox
•
Provides local or remote management
through Windows PowerShell or UI
•
Enables teams of up to 32 network
adapters
•
Aggregates bandwidth from multiple
network adapters whilst providing traffic
failover in the event of NIC outage
•
Includes multiple nodes: switch dependent
and independent
•
Multiple traffic distribution algorithms:
Hyper-V Switch Port, Hashing and
Dynamic Load Balancing
Virtual
adapters
Integrated Solution for
Resilient Virtual Machines
• Massive scalability with support for 64
physical nodes & 8,000 VMs
• VMs automatically failover & restart on
physical host outage
• Enhanced Cluster Shared Volumes
• Cluster VMs on SMB 3.0 Storage
• Dynamic Quorum & Witness
• Reduced AD dependencies
• Drain Roles – Maintenance Mode
• VM Drain on Shutdown
• VM Network Health Detection
• Enhanced Cluster Dashboard
Cluster Dynamic Quorum Configuration
Complete Flexibility for
Deploying App-Level HA
• Full support for running clustered
workloads on Hyper-V host cluster
• Guest Clusters that require shared storage
can utilize software iSCSI, Virtual FC or SMB
• Full support for Live Migration of Guest
Cluster Nodes
• Full Support for Dynamic Memory of Guest
Cluster Nodes
• Restart Priority, Possible & Preferred
Ownership, & AntiAffinityClassNames
help ensure optimal operation
Guest
Cluster
running
onona physical
Hyper-V
Cluster
node
restarts
failure
Guest
cluster
nodes
supported
with Livehost
Migration
Guest Clustering No Longer
Bound to Storage Topology
• VHDX files can be presented to multiple
VMs simultaneously, as shared storage
• VM sees shared virtual SAS disk
• Unrestricted number of VMs can
connect to a shared VHDX file
• Utilizes SCSI-persistent reservations
• VHDX can reside on a Cluster Shared
Volume on block storage, or on
File-based storage
• Supports both Dynamic and Fixed VHDX
Flexible choices for placement of Shared VHDX
Monitor Health of Applications
Inside Clustered VMs
• Upon service failure, Service Control
Manager inside guest will attempt to
restart the service
• After 3 failures, Cluster Service will
trigger event log entry 1250
• VM State = Application in VM Critical
• VM can be automatically restarted on the
same node
• Upon subsequent failure, VM can be failed
over and restarted on alternative node
• Extensible by Partners
Integrated Patching Solution
for Hyper-V Clusters
• Reduces server downtime and user
disruption by orchestration of cluster node
updates
• Maintains service availability without
impacting cluster quorum
• Detects required updates and moves
workloads off nodes for updates
• 2 Updating Modes: Self-Updating and
Remote Updating
• Uses Windows Update Agent or extensible
plug-in
U
Third-party plug-in for updates
Ensure Optimal VM Placement
and Restart Operations
• Failover Priority ensures certain VMs
start before others on the cluster
• Affinity rules allow VMs to reside on
certain hosts in the cluster
• AntiAffinityClassNames helps to keep
virtual machines apart on separate physical
cluster nodes
• AntiAffinityClassNames exposed
through VMM as Availability Set
Anti-Affinity
keeps
related
VMs
Upon
failover,
VMs
restart
order
Hyper-V
cluster
with
VMsinonpriority
eachapart
node
Replicate Hyper-V VMs from a
Primary to a Replica site
• Affordable in-box business continuity and
disaster recovery
• Configurable replication frequencies of 30
seconds, 5 minutes and 15 minutes
• Secure replication across network
• Agnostic of hardware on either site
• No need for other virtual machine
replication technologies
• Automatic handling of live migration
• Simple configuration and management
Primary Site
Replicated Changes
CSV on
Block
Storage
Secondary Site
SMB Share
File Based
Storage
Replicate to 3rd Location for
Extra Level of Resiliency
•
Once a VM has been successfully
replicated to the replica site, replica
can be replicated to a 3rd location
•
Chained Replication
•
Extended Replica contents match the
original replication contents
•
Extended Replica replication frequencies
can differ from original replica
•
Useful for scenarios such as SMB ->
Service Provider -> Service Provider DR
Site
Replication
canconfigured
be enabledfrom
on the
1st replica
to a 3rd site
Replication
primary
to secondary
Orchestrate protection and
recovery of private clouds
•
Protect important services by
coordinating replication and recovery of
VMM-managed private clouds
•
Automates replication of VMs within
clouds between sites
•
Hyper-V Replica provides replication,
orchestrated by Hyper-V Recovery
Manager
•
Can be used for planned, unplanned and
testing failover between sites
•
Integrate with scripts for customization of
recovery plans
VMs built on Optimized,
Software-Based Devices
Ease of Management & Operations
•
PXE boot from Optimized vNIC
•
Hot-Add CD/DVD Drive
Dynamic Storage
•
VMs have UEFI firmware with support
for GPT partitioned OS boot disks >2TB
•
Faster Boot from Virtual SCSI with Online
Resize & increased performance
Security
•
Removal of emulated devices reduces
attack surface
•
VM UEFI firmware supports Secure Boot
Enhancing VMConnect for
the Richest Experience
Improved VMBus Capabilities enable:
•
Audio over VMConnect
•
Copy & Paste between Host & Guest
•
Smart Card Redirection
•
Remote Desktop Over VMBus
Enabled for Hyper-V on both Server
& Client
Fully supports Live Migration of VMs
Enterprise-class.
Best-in-class performance and scale
for Microsoft workloads
Industry-leading support for
64-node clusters and 8,000 VMs per
cluster
64 TB VHDX virtual disk with online
resize for dynamically growing and
shrinking the VHDX file
Hyper-V Network Virtualization and
multi-tenant VPN gateway for inbox
software-defined networking (SDN)
solution
Simple and
cost-effective
Application
focused
User centric
SMB traffic offload to RDMA-capable
NICs for dramatic performance
improvements
Complete VM portability between
Windows Server and Windows Azure
without the need for VM conversion
Always-on remote access to corporate
information from trusted devices
Flexible guest clustering options for
file and block storage with shared
VHDX files that preserve dynamic
memory, live migration and storage
live migration for guest virtual
machines.
Same development model between
Windows Server and Windows Azure
Integration with Windows Azure
Backup for reliable and cost-effective
backup to the cloud.
VDI storage de-duplication
Common development tools between
Windows Server and Windows Azure
for a rich and complete environment
to build applications on-premises and
for the cloud
Consistent, inbox Windows experience
for remote users virtually anywhere on
any device
Unified application and device
management with common identity onpremises and in the cloud
Granular, policy-based data
protection and regulatory compliance
System Center Evolution
Management as
a Service
Cloud Application
Management
Application
Insight
Service
Management
IT Process
Automation
2012
App Controller
Operations
Manager
2010
2009
Service
Manager
Orchestrator
Server
2007
Virtualization
Virtual Machine
Manager
Anti-virus
Anti-Malware
Endpoint
Protection
Backup
2006
& Recovery
Data Protection
Manager
Monitoring 2000
Operations
Manager
1994
Configuration
Manager
Configuration
Enable the Client
Advisor
Enter the Data
Center
Integrated
Virtualization
Heterogeneous
Support
Embrace the Cloud
System Center Helps Deliver IT as a Service
Configure Deploy
App
Controller
Orchestrator
Private
Cloud
Virtual Machine
Manager
Operations
Manager
Configuration
Manager
Self-Service
App Owner
Service
Model
Service
Manager
Application Management
Service
Delivery and
Automation
Service
Manager
Public
Cloud
Virtualized
Data Protection
Manager
Physical
Monitor Operate
Service Delivery and Automation
Infrastructure Management
DC Admin
System Center 2012 Cloud and Data Center
Management
PRODUCTIVE
INFRASTRUCTURE
PREDICTABLE APPLICATIONS
YOUR CLOUD
Deliver flexible and cost effective
infrastructure with what you
already know and own
Applications power your business
Deliver predictable application service
levels with deep application insight
Private and public cloud computing
on your terms managed with
a common toolset
Heterogeneous support
Deep application monitoring and diagnosis
Flexibility with delegation and control
Process automation
Comprehensive application manageability
Applications self-service across clouds
Self-service infrastructure
Service-centric approach
Physical, virtual, and cloud management
Productive Infrastructure
Deliver flexible and cost-effective infrastructure using what you already know and own
Network
Compute
0%
Storage
Heterogeneous Support
50%
100%
0%
50%
100%
Compute
Fabric
Process Automation
Self-Service Infrastructure
Productive Infrastructure
Deliver flexible and cost-effective infrastructure using what you already know and own
Finance
Marketing
HR
Fabric
Heterogeneous Support
Process Automation
Self-Service Infrastructure
Predictable Applications
Applications power your business. Deliver predictable application service levels with deep application insight
Service Template — Multi-Tier .NET Application
Web
W2K8.VHD
OS Settings
MSDeploy package
and configuration
settings
App
App-V package and
configuration
settings
Networking
Service-Centric
Approach
Data
SQL DAC package and
configuration
settings
Load Balancer
Comprehensive Application
Manageability
Deep Application Monitoring
and Diagnosis
Your Cloud
Private and public cloud computing on your terms, managed with a common toolset
Hello, Jeremy (Self Service User)
Clouds
Virtual Machines
Flexibility with
Delegation and Control
Applications Self-Service Across Clouds
Physical, Virtual, and Cloud Management
Windows Azure
Pack
Customers
IT Admin
Windows Azure
Customers
IT Admin
In your datacenter
Customers
IT Admin
In your datacenter
Customers
IT Admin
Windows
Azure
Pack
Tenant experience
Homepage
Customer
ONE
Microsoft
Consistent
Platform
Service
Provider
Why choose System Center for Management
Infrastructure
Provisioning
Infrastructure
Monitoring
Automation
and Self-Service
Application
Performance
Monitoring
IT Service
Management
Enterprise-class
multi-tenant
infrastructure for
hybrid environments
Comprehensive
monitoring
of physical, virtual,
and cloud
infrastructure
Application-owner
agility while IT
retains control
Deep insight into
application health
Flexible service
delivery