© 2006 Cisco Systems, Inc. All rights reserved. Cisco ConfidentialPresentation_ID 1
Cisco DC TCO in a nutshell
Jose Moreno ([email protected])
Systems Engineer
12. April 2023
© 2006 Cisco Systems, Inc. All rights reserved. Cisco ConfidentialPresentation_ID 2
Improving the application performance
© 2006 Cisco Systems, Inc. All rights reserved. Cisco ConfidentialPresentation_ID 3
Cisco Application Networking Services
Wide Area Application Services
Consolidate IT infrastructure from the branch into the DC…
… and make sure your users do not suffer from performance degradation!
Data Center Application Services
Make sure your DC can scale in a flexible and cost-effective manner with Cisco ACE Load Balancers
Virtualizing your Load Balancing Infrastructure is crucial for the road to the "cloud"
Pay-as-you-grow, with license-based scalability models, that avoid costly fork-lift upgrades
© 2006 Cisco Systems, Inc. All rights reserved. Cisco ConfidentialPresentation_ID 4
Nexus technologies
© 2006 Cisco Systems, Inc. All rights reserved. Cisco ConfidentialPresentation_ID 5
Nexus Unified I/O
CNACNA
CNACNA
FC HBAFC HBA
FC HBAFC HBA
NICNIC
NICNIC
SAN (FC)
SAN (FC)
LAN (Ethernet)
LAN (Ethernet)
SAN (FCoE)
LAN (Ethernet)
CNA = Converged Network Adapter
SAN (FCoE) LAN (Ethernet)
Fewer interfaces means:Fewer cards pro Server power savings $$$Neater cabling Better airflow power savings $$$Fewer network switches $$$
© 2006 Cisco Systems, Inc. All rights reserved. Cisco ConfidentialPresentation_ID 6
Nexus Fabric Extender
Top-of-the-Rack (ToR) designs offer best cabling designs, but imply a lot of devices to manage
RackRac
kRac
kRac
kRac
kRac
kRac
kRac
k
A
Top of the Rack (ToR)
End of the Row (EoR)
Physical topology with Nexus FEX Logical topology with Nexus FEX
Optimized for cabling:
$$$
Optimized formanagement:
$$$
A1 A2 A3 A4
A
End-of-the-Row (EoR) designs are optimum from a network perspective: fewer devices to manage, but cabling is costly
© 2006 Cisco Systems, Inc. All rights reserved. Cisco ConfidentialPresentation_ID 7
Nexus VM-aware Networking
Non-Disruptive operational Model
Clear administrative borders
Mobility of network and security Properties
VMW ESX
Server 1
VMware vSwitch
VMW ESX
Server 2
VMware vSwitch
VMW ESX
Server 3
VMware vSwitch
VM #1
VM #4
VM #3
VM #2
VM #5
VM #8
VM #7
VM #6
VM #9
VM #12
VM #11
VM #10
VEM VEM VEM
© 2006 Cisco Systems, Inc. All rights reserved. Cisco ConfidentialPresentation_ID 8
Achieving synergies
© 2006 Cisco Systems, Inc. All rights reserved. Cisco ConfidentialPresentation_ID 9
I/O Consolidation + Fabric Extender
Multiple points of managementFC
Ethernet
Blade switches
High cable count
Unified Fabric with Fabric extender Single point of management Reduced cables
Fiber between racks
Copper in racks
Traditional switching FCoE + Fabric Extender
© 2006 Cisco Systems, Inc. All rights reserved. Cisco ConfidentialPresentation_ID 10
I/O Consolidation and VM-aware networking
Hypervisor
VirtualMachineVirtual
MachineVirtual
MachineVirtual
MachineVirtual
MachineVirtual
Machine
Reduced CPU utilization through hypervisor bypass (VMDirectPath technology)
vhbas supported!
Standards-based (SR-IOV)
More VMs per server
VM-aware networking for FC and Ethernet
Hypervisor
VirtualMachineVirtual
MachineVirtual
MachineVirtual
MachineVirtual
MachineVirtual
Machine
© 2006 Cisco Systems, Inc. All rights reserved. Cisco ConfidentialPresentation_ID 11
The architecture with the whole picture:
Unified Computing
© 2006 Cisco Systems, Inc. All rights reserved. Cisco ConfidentialPresentation_ID 12
Reduced Points of Management
Universal Interconnect
Scalable Infrastructure Optimized Virtualization Scalable Virtualization
Unified embedded management and service processor
Unified Fabric to PCIe bus
Fabric Extender for wire once connectivity
2 to 640 4 core processors per virtualization pool (Intel Nehalem x86 chipset)
up to 6000 Virtual Machines in a virtualization pool
Open interfaces accommodating existing management investments
Seamless connectivity to DCE, FCoE, & 10GE
Up to 320 nodes managed as a single integrated system
Network Optimized for granular VMotion and dynamic provisioning control
up to 384 GB per node for in memory computations
Site TCO (CAPEX and OPEX)
1.Reduced ‘System’ Power2.Lower Cooling3.Better Use of Space4.Lower Power/Site
Organization TCO (OPEX)1.Fewer FTE/”Service”2.Faster Provisioning3.Seamless Repurposing4.Better Coordination
Platform TCO (CAPEX and OPEX)
1.Radically Fewer Components2.Lower HW/SW Costs3.More VM’s Per Node4.Better Performance Per Node
35%15%20%
The “Unified Computing” Solution CAPEX and OPEX Solution Targets
Tight integration: Unified Computing System
© 2006 Cisco Systems, Inc. All rights reserved. Cisco ConfidentialPresentation_ID 13