What is coming for VMware vSphere?
Delivered at VMUG DK/UK/BE in November 2014. Session is all about vSphere futures, what can be expected in the near future.
2. • New project New Server
• Consolidation of apps
• Operational complexity
From Physical
3. • Faster time to market
• Standardization a must
• New operational models
To Virtual
4. • Automation / Orchestration
o Daily tasks
o Monitoring
o Configuration
• Consistency
• Repeatable
Operations
5. Disclaimer
• This presentation may contain product features that are
currently under development.
• This overview of new technology represents no commitment
from VMware to deliver these features in any generally
available product.
• Features are subject to change, and must not be included in
contracts, purchase orders, or sales agreements of any kind.
• Technical feasibility and market demand will affect final
delivery.
• Pricing and packaging for any new technologies or features
discussed or presented have not been determined.
• All of the features and futures discussed have been presented
at VMworld.
7. Shared-All virtual machines
through forking aka VMFork!
Benefits:
o Pre-emptive memory sharing
o Lower CPU – avoid boots, logins,
app launches
o I/O reduction: Boots / App-
Launches generate I/O that our
forking design can avoid.
Project Fargo
Copy On Write
Shared
memory
Unique
memory
Base
Disk
Delta
Disk
Copy On Write
Quiesced
Parent
VM
Child
VM
8. Fork 1 Fork 2 Fork 3 Fork N
Just in time
Parent
VM
Load Balancer
Hello World!
500ms ago
this VM
didn’t exist.
By the time
you read this,
the VM that
created this
page has
already
been
destroyed.
Hello World!
500ms ago
this VM
didn’t exist.
By the time
you read this,
the VM that
created this
page has
already
been
destroyed.
Hello World!
500ms ago
this VM
didn’t exist.
By the time
you read this,
the VM that
created this
page has
already
been
destroyed.
VMware vSphere
Hello World!
500ms ago
this VM
didn’t exist.
By the time
you read this,
the VM that
created this
page has
already
been
destroyed.
10. • Scale out architecture
• Built into the hypervisor
• Aggregates host local
storage
• Dynamic scalability
• Flash optimized
• VM-centric data operations
10
Virtual SAN
11. • All Flash Virtual SAN
configurations
o Enterprise MLC for write buffering
and consumer grade for reads
• Virsto file-system integration
o Fast cloning
o New efficient snapshotting
mechanism
• JBOD support
o Useful for blade environments!
• 64 Node clusters
11
Enhancements
12. Offerings
VMware EVO:RAIL
Hyper-Converged
Infra Appliance
• Maximum simplicity
• Prebuilt Hyper-Converged
Infrastructure Appliance
(HCIA)
• EVO:RAIL Rapid
Deployment and
Configuration Engine
• Single SKU for hardware,
software, and SnS
• Transformational user
experience
Virtual SAN built from
HCL components
• Maximum choice
• Go through HCL to
find certified
combination
• Assemble hardware
• Procure many SKUs:
• Hardware
platform
• Hardware SnS
• Software + SnS
• Possible services
Virtual SAN
Ready Node
• Manufacturer specifies
hardware configuration
from HCL components
• Hardware is prebuilt;
customer installs and
configures vSphere &
Virtual SAN
• Procure:
• Hardware platform
• Hardware SnS
• Software + SnS
• Possible services
13. EVO:RAIL
• Builds on trusted
technology:
o vSphere + Virtual SAN + Log Insight
+ EVO:RAIL Engine
• Prescribed 2U/4N platform
from qualified partners
• Automated scale out up
to 4 HCIAs
• Time to value to first VM as
little as 15 minutes
• Non disruptive patch &
upgrade
Hyper-Converged Infrastructure
Appliance (HCIA)
x4
x3
x2
x1
TOR Switch
14. User Experience
• Single SKU / single point of
contact
• VMware EVO:RAIL Engine
• Rapid configuration in under
15 minutes
• Simple management
• One-click non-disruptive
upgrades
• Automatically scales out
Deploy
Configure
Manage
16. EVO:RACK
• Pre-qualified Data Center-
Scale Hyper-Converged
Cloud Infrastructure from
Partners
• Management:
o Simplified SDDC deployment,
configuration, and provisioning
o SDDC lifecycle – non-disruptive
patching/upgrading
o SDDC-wide capacity planning
o Integrated management of both
logical and physical resources
o Configuration and provisioning of
end-to-end network infrastructure
• SDDC up and running in
under 2 hours
17. EVO:RACK
Solution
Virtualized Infrastructure
Appliance
Data Center-Scale
Cloud Infrastructure
VMware
Software
vSphere, Virtual SAN, EVO:RAIL
Engine
Full vCloud Suite, Virtual SAN, NSX,
Integrated Virtual + Physical
Networking, EVO:RACK Software
Managed
Hardware
Server + Built-in Storage
Server + ToR Leaf/Spine Switch
+ JBOD or DAS Storage
Server
Specifications
2U / 4N Appliance; ability to stack
appliances
Rack-mounted servers;
Deployment
Scalability
Up to 4 appliances Multiple racks
Key Use Cases ROBO, VDI IaaS, VDI
18. User Experience
• Pre-configured rack(s) arrive
at a customer site with
o Pre-integrated SDDC software
suite, EVO:RACK software and ToR
switches
• Customers:
o Connect rack utilities: network
uplinks & power
o provide network info IP range, DNS,
NTP…
• EVO:RACK: auto-provision IP,
L3 fabric
• EVO:RACK: single point of
control for operations &
lifecycle management
Single Virtual Rack
(ESX, Virtual SAN, NSX)
19. Old World!
Storage Consumer Challenges:
o Lengthy provisioning cycles
o Difficult to make adjustments
o Lack of granular control
o Complex troubleshooting
o Frequent data migrations
Storage Provider Challenges:
o Fragmented device management
o Rigid capacity and data services allocation
o Complex LUN/Volume management
VMware vSphere
Storage Array - A Storage Array - B
20. • Virtual disks are natively
represented on arrays
• Enables VM granular
storage operations
• Integrates with Storage
Policy-Based
Management
• Supports existing
storage I/O protocols
(FC, iSCSI, NFS)
• Ecosystem-wide
initiative
20
Virtual Volumes
vSphere
admin
Storage
admin
21. With VVOLs
Storage Policy Based MGMT
VMware vSphere
Virtual Datastore
Virtual Datastore
Virtual Datastore
Storage Policy
Capacity
Availability
Published Capabilities
Snapshot
Replication
Dedupe
Performance
Data Protection
Security
container
Storage Policy Based MGMT
o Policy driven, VM-centric
control plane
o Dynamic composition of
storage services
o Intelligent placement and
transparent remediation
o Common across
heterogeneous devices
Virtual Volumes
o Virtual disks natively
represented on external storage
o Granular control of native array
data services on a per VM basis
o No more LUNs/Volumes
Virtual Datastore
container
container
container
Storage Array - A Storage Array - B
22. IO Filters
• Integration API for
Software-based Data
Services within vSphere
• Performance of kernel-mode
with the stability of
user-mode
• Tight integration with
vSphere and Storage
Policy Based
Management
• Granular provisioning
and control on a per VM
basis
Storage Policy Based MGMT
VMware vSphere
Replication
Filter
Caching
Filter
Storage Array - A Storage Array - B
Storage Policy
Performance
Data Protection
24. SIOC
• VM IOPS reservations
• Queue depth allocation of
each host depends on the
VM reservation
• Migrate workload if
reservations are not met
• Balance reserved IOPS
usage in cluster
• Reservations as VM
placement constraints
24
IOPS Reservations
R: 300 R: 200 R: 500
Storage DRS Datastore Cluster
25. • Aware of deduplication using VASA 2.0
o Storage DRS manages logical space while keeping virtual disks in
the same dedupe pool
• Deep integration with vSphere Replication
o Storage DRS understands space usage of replica disks
• Storage DRS identifies replicated datastores
• All recommendations are in sync with replication
policies:
o Automated moves within the same consistency group
o Manual moves for all VMs residing on replicated datastores
o Accounting of replication overhead due to Storage vMotion
• Aware of thin provisioned VMs and Datastores
25
More SDRS
26. Problem:
• Host has a loss of storage
connectivity
o APD: All Paths Down
o PDL: Permanent Device Loss
o Difficult to manage VMs running on
APD/PDL affected hosts
Approach:
• VMs are restarted on healthy
hosts
26
vSphere HA
All Paths Down
e.g. Path down
or port disabled
Permanent Device Loss
e.g. Array misconfiguration,
Host removed from
LUN’s Storage Group
28. • FT support for up to 4 vCPUs
• New fast check-pointing
technology to keep primary
and secondary in sync
• Protect mission-critical, high-performance
applications
regardless of OS
• Continuous availability: zero
downtime, zero data loss for
infrastructure failures
• Fully automated response
28
vSMP FT
Instantaneous Failover
Fast Checkpointing
10Gbps Infra
29. New Architecture
• No more vLockstep
o SMP FT protocol
• No more shared VMDKs
o Both primary / secondary have own VMDK!
• 10Gbps networking is a requirement!
31. • vMotion across vCenter
Server boundaries
• Easily move VMs across
(distributed) vSwitches,
folders and data centers
• Builds upon vMotion without
shared storage
• Simplifies vCenter Migration
and consolidation
• Aligns vMotion capabilities
with larger data center
environments
• Leverage VVOL capabilities
31
vMotion Anywhere
32. • Targeting cross-continental
US distances –
up to 100 ms RTTs
• Maintain standard
vMotion guarantees
• Leader in VM flexibility
• Benefits
o Permanent migrations between
data centers
o Disaster avoidance (DA)
o SRM/DA testing
o Multi-site load balancing
o Follow the sun
32
Long Distance
34. Metric Windows Appliance*
Hosts per VC 1k ✔
Powered on VMs per
10k ✔
VC
Hosts per cluster 64 ✔
VMs per cluster 6k ✔
Linked Mode 10 ✔
* Single instance VC with embedded vPostgres
Equal
35. • Web Client Usability
o Tasks back in bottom
o Right click flat menus!
• Linux Appliance with performance/feature parity
with Windows (e.g., scale, federation)
• Linked Mode vs. SSO converging functionality
• Performance Improvements
o vCenter task concurrency, performance charts, web client, etc.
35
Usability
36. • Content Library provides
storage and versioning of files
including VM templates, ISOs,
and OVFs.
• Includes powerful publish and
subscribe features to
replicate content
• Simplifies management of
templates
• Enables synchronization
between local and remote
vCenter Servers
Content Library