During our on demand webinar, “Simplifying the Large-Scale Hybrid Cloud”, Storage Switzerland and Axellio discuss how Microsoft Azure Stack HCI and Axellio’s FabricXpress Servers can deliver new levels of consolidation in the enterprise. Learn how to intelligently leverage Azure to simplify operations like data protection, business continuity, and data center operations – while deploying less infrastructure and less software for your demanding on-premises workloads.
3. What Makes Hybrid Cloud So Difficult
(why does it need simplification?)
Hybrid typically means either “here” or “there”
• Application is either running on-premises or in the cloud
• Moving the applications data is difficult
• Move data with application
• Continually replicate data to cloud
• Application typically needs conversion between
on-premises hypervisor and cloud hypervisor
• As a result applications are moved infrequently
4. How To Simplify
Enterprise Size
Hybrid Cloud
● Simplify and Consolidate
On-Premises
● Leverage Cloud Services
to work with On-Premises
● Create a Path for Workload
Migration
5. Simplify On-Premises
● Hyper-converged simplifies by
● Converging compute, storage and networking
under the hypervisor
● Traditional Hyper-converged though does
NOT Scale Efficiently
● Typically workload specific
● Not typically used for storage intensive
workloads
● Can’t handle both scale-up & scale out
applications in the same cluster
● Complexity re-introduced as HCI scales
● Just add a node adds complexity
6. Consolidate On-Premises – The Hardware Matters
More Powerful, Better
Equipped Nodes
• NVMe Flash Nodes
• High Performance Compute
Leads to
• Ability to run all workloads,
including storage intensive
• Limits node sprawl
7. ● Leverage services first
● Backup and recovery
● Cluster Quorum
● Monitoring
Simplifying the Transition to Hybrid Cloud:
Step 1
8. ● Moving Applications
● Simplify by using
compatible infrastructures
(Azure)
Simplifying the Transition to Hybrid Cloud:
Step 2
10. HYBRID CLOUD ADOPTION:
HOW TO DEPLOY
REPLACE AGING INFRASTRUCTURE
WITH SMALL CLUSTERS OF LARGE
SERVERS
EXTEND TO AZURE WITH SIMPLIFIED
CLOUD MANAGEMENT & SECURITY
SERVICES
CHOOSE A VALIDATED SOLUTION
FOR EASE OF DEPLOYMENT &
MANAGEMENT
• FabricXpress’ unique technology allows you
to scale up and out in the same cluster
• Up to a petabyte of NVMe all-flash storage
in just 6U scaling to 4 petabytes in just 18U
• Consolidate SQL Databases*, Data
Warehouse, File Servers, CRM, ERP into a
single powerful HCI cluster
• The only proven solution that drives enough
IO to support all workloads in one HCI
cluster
• Axellio takes validated to the next level with
an end to end optimized solution from pilot
to support
• FabricXpress, using off the shelf
components in a unique way to push the
performance of the solution, is validated
against rigorous Microsoft testing
• Microsoft has created a standard in
hypervisor from on-premises to cloud for
simple and familiar management
• With built-in integration to Azure from
Windows Admin Center – you can easily
integrate on-premises workload with:
• Azure Site Recovery & Backup
• Cloud Witness
• Azure Monitor
• And more…
*request a copy of benchmark testing of multi-application workloads performed in same cluster
11. • An “Inside-Out” approach to Hybrid Cloud
• Deployed as an Enterprise HCI solution for
on-premises and edge that easily extends to
Azure for hybrid capabilities
• Built on a server that is designed to support
workloads in a new way
FABRICXPRESS FOR AZURE STACK HCI:
A SIMPLE VALIDATED SOLUTION
SIMPLICITY | SCALABILITY | PERFORMANCE
13. Benefit of Large Servers:
• Traditional HCI clusters for single use workload (i.e. VDI only)
aren’t your only option:
• More available NVMe devices allows for larger CPUs to be fully utilized
= capability to carve out larger VMs
• With ability to serve up larger VMs comes the ability to run
workloads concurrently within the same virtualized environment
without QoS configuration
Benefit of Smaller Clusters:
• Greater Simplicity on top of the already simple HCI
approach
• Fewer servers and network ports
• Less socket based licensing costs
• Larger CPU and available capacity per node enables greater space
and power efficiency
SMALL CLUSTERS OF LARGE SERVERS:
A SIMPLE & POWERFUL APPROACH
Up to 176 Processor Cores
352x Hyper Threaded Cores
Up to 4 TB RAM
Up to 920 TB Raw Capacity*
Capable of up to 1M IOPS per node
@350us latency
HP SN6000B 16Gb FC Switch
47434642454144403935383437333632312730262925282423192218211720161511141013912873625140
HP SN6000B 16Gb FC Switch
47434642454144403935383437333632312730262925282423192218211720161511141013912873625140
IN JUST
6U
*Increased capacity supported as higher capacity SSDs become available
14. FABRICXPRESS FOR AZURE STACK HCI:
DESIGNED FOR ALL WORKLOADS
DATABASES TRADITIONAL VM CONTAINERS
3 PERFECT WORKLOADS FOR FABRICXPRESS
• Legacy Business Apps
• Scale-Out
• Fewer instances with large
storage IO requirements
OLTP | OLAP
• Standard Business Apps
• Scale-Up
• Many instances taxing VM
resources
VDI | File, Print, & Web
Servers | Active Directory
• Modern Business Apps
• Scale-Out
• On-prem with cloud expansion
capabilities
Cloud Native | Kubernetes
| Docker
15. WHEN TO TURN TO AXELLIO
You’re trying to
driving datacenter
consolidation in your
next refresh due to
budget, sprawl, etc…
Looking at HCI but
have not found a
solution with enough
storage and/or
compute to the make
move
You’re looking for
expertise to help get
you into a hybrid cloud
with simplicity,
scalability, and
performance
17. Simplifying the Enterprise Hybrid Cloud
with Azure Stack HCI
For complete audio and Q&A please register for the
On Demand Version: bit.ly/SimplifyCloud
Editor's Notes
More available NVMe devices allows for larger CPUs to be fully utilized vs.
***448 post data protection before dedup and compression – which we quote around 2.5 on average…
SAN vs. HCI (a scale-out technology)
Scale-Up applications have not typically been candidates for HCI which forces you away from HCI for your scale up applications
Axellio has changed this – proving that a scale up data warehouse in a FX-HCI cluster can perform at the same rate as a SAN and still run the rest of your VMs
By enabling a larger VM you can present a large enough virtualized server to support these scale-up applications
have additional resources left over for additional workload concurrency