1. UCS Roadmap Name

2. UCS Innovations for Converged Solutions
Infrastructure Management
UCS Director
UCS “All-in-One” Solution
Virtual Interface Card
VM-FEX for ESX and Hyper-V, 10GBaseT, 40Gb, Ethernet, USNic, RoCE
C-Series Rack Integration
Single Wire Management
Next-Gen Servers
Ivy Bridge, Brickland & Grantley
UCS Central
Multi-UCS Domain Manager
Density Optimized Solutions

3. Post-Grantley Consolidated UCS Compute Portfolio Performance Optimized for Bare Metal, Virtualized, and Cloud Applications
Cisco UCS: Many Server Form Factors, One System
Industry-Leading Compute Without Compromise
Scale Out
Enterprise Performance
Intensive/Mission Critical
UCS C240 M4
Ideal Platform for Big Data, ERP, and Database Applications
UCS C460 M4
Mission-Critical, 4-Socket Server for Large, CPU-Intensive Applications
Rack
UCS C220 M4
Versatile, General Purpose Enterprise Infrastructure, and Application Server
UCS B460 M4
Mission-Critical, 4-Socket Blade for Large, CPU-Intensive Bare Metal and Virtualized Applications
UCS M-Series Modular Servers
Modular servers optimized for Cloud-scale deployments
UCS B260 M4
Mission-Critical, 2-Socket Blade for Large, CPU-Intensive Bare Metal and Virtualized Applications
UCS B200 M4
Optimal Choice for VDI, Private Cloud, or Dense Virtualization/ Consolidation Workloads
Blade
UCS B420 M4
Enterprise Class, 4-Socket Blade for Large, Memory-Intensive Bare Metal and Virtualized Applications

4. UCS Compute Rack Roadmap
Q4CY 2013
Q1CY 2014
Q2CY 2014
Q3CY 2014
Q4CY 2014
Q1CY 2015
Q2CY 2015
Q3CY 2015
Q4CY 2015
C460 M4
Expandable Servers (4S –EX)
Brickland / Haswell
C460 M4
C260 M2
Enterprise Class Server (4S –EP)
C420 M3
Next Gen 2S –EP Rack (2U)
Enterprise Class Server (2S –EP)
Grantley
BDW Refresh
C240 M3
Next Gen 2S –EP Rack (1U)
Enterprise Class Server (2S –EP)
Grantley
BDW Refresh
C220 M3
Value Server (2S –EN)
C24 M3
Value Server (2S –EN)
C22 M3

5. UCS Mini Topology
Integrated server, networking and management in a single chassis
New Fabric Interconnects fit in current IOM bays
Same compute blades, fans, power supplies
Fully capable UCS Manager
Connectivity to external storage
At FCS: iSCSI, FCoE
Post FCS: FC
Service Profiles, UCS Central, XML APIs, etc…
Leverages UCS Central for remote site management at scale
Connect up to 2 Chassis, 20 servers
16 blade servers 7 rack mount servers C220M3 and C240M3

6. UCS M-Series Modular Servers
True Server Disaggregation
Compact Chassis
8 Compute cartridges
Based on Cisco System Link Technology
3rd Gen VIC extends UCS fabric to within the server
Shared Local Resources
Four shared SSDs in the chassis
Shared dual 40Gb connectivity
Compute Density
16 Intel Xeon E3 Compute nodes in 2RU chassis
Each cartridge holds two independent compute nodes
Lightweight Compute Cartridge
Two Independent Intel Xeon E3 Servers
No adapters or HDDs
Key Elements:
Based on System link and targeted for Cloud scale applications
8 compute slots, “First” cartridge is a E3
M-Series modular server is an extension to the current UCS portfolio. The platform is targeted for specific workloads and not a general purpose compute platform.  The platform consists of a new chassis in a 2 RU form factor. It houses 8 front loadable cartridges. Each cartridge in turn comprises of 2 independent server nodes. All of the servers in the chassis share the peripherals. The back of the chassis can hold up to 4 SSDs, 2 power supplies and 2 * 40 Gb network uplinks. A  single chassis can hence have  up to 16 servers, 1/2 TB of Memory and up to 6.4 TB of disk capacity. Of course, most of the deployments will  have multiple chassis or multiple racks.
Shared Local Resources
Network and storage resources

7. UCS Management Portfolio
UCS Director API
UCS Director
Policy Driven, Application Centric Infrastructure Management and Orchestration
Non-Cisco Infrastructure
API
UCS Central
Policy Driven Multi DC, Multi-Domain Management
Virtual Machines
API
API
Network Devices
UCS Manager Domain 1
UCS Manager Domain x
API
CIMC
Storage
FlexPod
Stand-Alone UCS C-Series
Integrated & Converged
Infrastructure
Unified Computing System
Servers
Basic Management Functionality
Advanced Infrastructure Abstraction & Automation

8. Cisco UCS Director Turn-Key Solution
Available
Cisco UCS Director Turn-Key Solution
Secure Cloud Container
VMs
Compute
Network
Storage
On-Demand
Automated Delivery
Domain Managers
OS and Virtual Machines VM VM
Bare Metal
Policy-Driven Provisioning
Virtualized and Bare-Metal
Single Pane of Glass End-to-End Automation and
Lifecycle Management
Compute
Compute and Hypervisor
Network
UCS Director A B C
Network and Services
Storage
Tenant A
Tenant B
Tenant C

9. Common Converged Infrastructure Customer Questions
My application is slow. Is it UCS? Network? Storage?
Do I need to buy more capacity, or can I just adjust how we’re using it?
How do I view UCS fabric utilization?
Typical Management Toolset
Windows SCOM
Linux Nagios and Cacti
VMware vCenter
Typical Management Toolset
Cisco UCS Manager
Cisco Prime Data Center Network Manager
NetApp OnCommand
Can I reconfigure to use more of the capacity we have?
Can we simplify the view of all the converged infrastructure pieces?

10. UCS Performance Manager – Application view: physical/logical/virtual resources
Define application as a set of operating systems
Dynamic view shows which UCS components are supporting those servers
Event rainbow shows components affected by fault issues or resource constraints
Groups
Devices
vSphere Endpoints
vSphere Hosts
UCS Service Profiles
UCS Blades
UCS Chassis
UCS IO Cards

12. Cisco ACI Better Together with NetApp
December 2014

13. A Network. Not a Network Emulation
Simplified, Single Architecture performing both overlay and underlay functions
Provides benefits of SDN + Policy without the complexity of a separate overlay and underlay
Reduces Complexity and adds operational simplicity
Distributed fabric Intelligence reduces engineering
Pre-Architected, Pre-Validated, Pre-Hardened
Optimal Traffic forwarding
Location Independent Forwarding
Congestion Monitoring
Flowlet Switching
Fabric Load Balancing
Full Real-time Visibility
Tenent Level
Network Level
Application Level
Atomic Counters
APIC
Key Messages
- ACI is a combination of Cisco Nexus 9000 switches and the Application Centric Infrastructure Controller (APIC)
- It provides a network fabric that is self assembling and policy driven
- Switching and routing are a result of defined policy, not explicit configuration
- Provides the benefits of integrated overlay networking without compromising performance, visibility or troubleshooting capability (unlike VMware)
Full Visibility of Overlay and Underlay for telemetry and troubleshooting

14. Software Define Infrastructure based on the Application Requirements
Simple, Scalable and Fast
Common Policy Model
Abstracted Policy model based on application requirements
Declarative model based on the scalable control of intelligent objects
Infrastructure operates as single system providing specific connectivity and services based on Application definitions
Allows application developers to succinctly and easily describe Infrastructure as Code
Policy allows Infrastructure and Dev Teams to use common requirements language to accelerate application deployment
Every Software and Hardware component is a programmatic object
Policy integration northbound with automation toolsets (e.g Openstack) and southbound with 3rd party network services vendors (e.g F5)
APPLICATION
NETWORK
PROFILE
EPG
EPG
EPG
f/w
ADC
ADC
WEB
app db
Key Messages
- The ACI policy model abstract away the complexity of manually stitching together network connectivity.
- The declarative nature of the model enables administrators to focus on “WHAT” they want to achieve rather than the syntax of “HOW”
- Fully API driven to enable third party programming and DevOps models via tools such as Chef and Puppet
- Application specific policies enable network visibility, metric, diagnostics and troubleshooting at the application level, not just switch ports.
Contract
Contract
Software Define Infrastructure based on the Application Requirements

15. Why ACI? The data center is both Virtual and Physical
Enterprise Scale and Performance requires hardware acceleration
A SINGLE architecture to deliver performance, programmability, agility and reduced complexity
An Application Centric Policy Model that dynamically defines the network fabric by mean of the application requirements
An AUTOMATED network fabric for both virtual and physical workloads and services

16. Traditional Switching Systems + ASICs+ Software
Why the Cisco Approach
ACI
Common Policy Model
Physical + Virtual
Secure Workload Placement
Virtual Overlay
Application Visibility + Health Metrics
SIMPLICITY, SCALE , SECURITY
Scale-out Performance
Traditional Switching
VM Mobility
Full Network Automation
Key Messages
- ACI is a far more comprehensive solution than “SDN”.
- It provides the benefits of programmability, flexibility, and integrated visibility that overlays alone can’t provide.
- The Merchant+ strategy of the Cisco Nexus 9000 family of switches enables low cost networking while still providing best in class functionality
- ACI is an open architecture designed to enable a broad ecosystem that is agnostic to the hypervisor vendor or third party L4-7 appliances
A Network, not a Network Emulation
Integrated Hardware and Switching Software
SDN LAN Emulation
“DIY” Basic Switching
Open Architecture
VM-Based Policy
White Box
Merchant Silicon
Software Only
Systems + ASICs+ Software
Lower TCO

17. Cisco ACI Based FlexPod
New
Cisco ACI Based FlexPod
Application
Catalog
Web
App DB
Compute
Network
Storage
UCS Director
Network
Compute
Storage
Web
Stateful FW
Appl.
L4-L7 services
Stateless FW
Load Balancing DB
Connection to Application
Web
Cache locally
Appl.
Primary Storage
IOPS
Mirrored
Archived DB
2 High IOPS
Data Mirroring
Web
2 VMs, High IO
Appl.
8 VMs, 10 GB DB
2 Baremetal
Key Messages
- New Cisco Validated Design (CVD) provides and architecture for UCS, ACI and NetApp
- Covers integration with Cisco UCS Director for IaaS/private cloud deployments
APIC
Clustered Data ONTAP
UCS Manager

18. Cisco ACI Takeaways Cisco Application Centric Infrastructure SPEED
Fixed Workloads
Variable Workloads
SPEED
NETWORK and SERVICES Delivered in minutes
SECURITY
INHERENT Security and INTEGRATION
TELEMETRY
Rich TELEMETRY & Application HEALTH SCORE
POLICY
Policy-based deployment/governance
Physical & virtual
Key Messages
- Cisco ACI decreases time to delivery of network services (L2-7) for physical and virtual workloads
- The ACI increases security through configuration consistency, integrated policy based ACLs (contracts), and integration with security services (IDS/IPS, FW, etc…)
- Visibility, at the application level, of network performance enables a higher level of service assurance correlated back to applications.
- The policy model reduces complexity, decreases downtime, simplifies troubleshooting and auditing of the network and applications
OPEN and AGNOSTIC

20. Appendix – Additional Detail
Appendix – Detail Slides

21. M4 Scalable EX Blade Highlights
Designed for the most processor, memory, and IO-intensive workloads
Provides mission-critical RAS features
Offers the benefits of blade density and ecosystem with the added flexibility of a truly scalable system
UCS B260 M4
Full-Width, 2-Socket (Brickland EX – up to 30 cores), 48 DIMMs
Up to 320 Gb of IO Bandwidth Optimized for Cisco VIC
Up to 6 TB of memory Up to 96 DIMM sockets *64GB DIMMs available Post-FCS
Optimized for both 2 and 4 socket scale (w/ required Scalability connector)
Four E7v2 CPUs
Up to 60 Xeon cores
UCS B460 M4
Double-High -Full-Width, 4-Socket (Brickland EX – up to 60 cores), 96 DIMMs
Dual / Quad Socket EX scalable design

22. Next Gen Density-Optimized Blade Server
An uncompromised combination of CPU, Memory, IO, and expansion capabilities in a modular blade form-factor
Designed for a wide range of enterprise workloads including virtualization and bare metal applications
Provides enterprise-level capabilities and features
Ultimate density-optimized general compute blade platform
Two E5v3 CPUs
Up to 18 cores per socket
24 DDR4 DIMM Slots
Up to 2133 MHz speeds
Flexible Storage
Controller / Cache / Media Options
UCS B200 M4 Density-Optimized
Enterprise Blade Server
Up to 80G of I/O
One PCIe 3.0 mezzanine slot
Modular LOM (mLOM)
3rd Generation VIC

23. Next Gen Dense 4-Socket Blade Server
CQ1’15
Next Gen Dense 4-Socket Blade Server
A density-conscious quad-socket blend of CPU, Memory, Storage, IO and expansion capabilities in a full blade form-factor
Designed for a wide range of enterprise workloads including virtualization and bare metal applications
Provides enterprise-level capabilities and features
Ultimate 4-socket density-optimized enterprise blade platform
UCS B420 M4 Density-Optimized
Enterprise Blade Server
Four E5v3 CPUs
Up to 18 cores per socket
48 DDR4 DIMM Slots
Up to 2133 MHz speeds
Flexible Storage
Controller / Cache / Media Options
Up to 160G of I/O
Three PCIe 3.0 mezzanine slot
Modular LOM (mLOM)
3rd Generation VIC

24. UCS “All-in-One” Solution (aka “Mini-UCS”)
UCS innovation optimized for Branch Office and Remote Sites
Integrated server, networking and management in a single chassis
New Fabric Interconnects fit in current IOM bays
Same compute blades, fans, power supplies
New 110v AC PSU option
Fully capable UCS Manager
Connectivity to external storage
At FCS: iSCSI, FCoE
Post FCS: FC
Service Profiles, UCS Central, XML APIs, etc…
Simple to setup, leverages UCS Central for remote site management at scale

25. UCS Manager 3.0(1) for UCS Mini
Available
UCS Manager 3.0(1) for UCS Mini
UCS Manager 3.0(1) is a UCS Manager release that only supports UCS Mini
6324 Fabric Interconnect only
Support for updated 5108 chassis only (UCS-5108-AC2 & UCS-5108-DC2)
UCS Manager 3.0(1) is based on UCS Manager 2.2(1d)
Does not include features and platforms supported in UCS Manager 2.2(2c) and later
New features:
Dual Line Power Supply Unit supporting 110V
USB support
QSFP+ (Scalability Port) Breakout Support
Licensing Support for QSFP+
Power Capping
Unsupported features in this release:
Ethernet Switching Mode
FC End Host Mode
Private VLANs
Port Security
KVM Virtualization (not explicitly disabled)

26. Cloud-Scale Inverts Computing Architecture
Core Enterprise Workloads
Cloud Scale
SCM
Legacy
CRM
Online Content
Mobile
IoT
ERP/Financial
Gaming
eCommerce
Many Applications
Server
Single Server
Hypervisor
Many Servers
Single Application
Key Elements:
Focused on making enterprises efficient /agile and virtualization + HW resiliency has been the right tool, right sizing by virtualization
Cloud need elastic scaling, resiliency at scale and the tool of choice is horizontally scaling app
The new generation applications can be flexibly sized. Right sizing by app instance

27. Cisco UCS Performance Manager
UCS Integrated Infrastructure
VMs
Compute
Network
Storage
Automated Discovery
OS and Virtual Machines
Storage
Network
Compute
Virtualized and Bare-Metal
Compute and Hypervisor
Network devices VM
Bare Metal
Out of the box monitoring
rules and thresholds
Single Customizable Solution for UCS Infrastructure Monitoring
UCS Performance Manager

28. Extend beyond the UCS Fabric
Determine if workloads and applications are affected by UCS configuration or capacity constraints
Compare current and historical performance across converged infrastructure components to find constraints
Key point: This typically requires multiple, independent tools