You are on page 1of 93

BRKDCT-2218

Data Center Network Design


for Midsize Enterprises

Fausto Vaninetti, Data Center


EMEAR DAT Consultant Systems Engineer
Cisco Spark
Questions?
Use Cisco Spark to communicate
with the speaker after the session

How
1. Find this session in the Cisco Live Mobile App
2. Click “Join the Discussion”
3. Install Spark or go directly to the space
4. Enter messages/questions in the space

cs.co/ciscolivebot#BRKDCT-2218

© 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public
…until you get to something
you’re happy with

Start small

… then grow
EEE education

Hint: leverage yet expand your skill set (even in IT)


BRKDCT-2218
exposure experience

© 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 4
Agenda

• Introduction
• Beyond STP, from Networks to Fabrics
• Scaling Brownfield and Greenfield Data Center Networks
• Data Center Interconnect Solutions
• Feature-Specific Considerations
• Programmability, Automation & Orchestration
LOTS of Related Sessions, My Recommended List
Session ID Title Presenter Date / Time

BRKACI-2125 ACI Multi-Site Architecture and Deployment Max Ardica Check on CiscoLive App

BRKDCT-2949 Building Data Center networks with VXLAN BGP- Lukas Krattiger Check on CiscoLive App
EVPN Part I
BRKDCT-3378 Building Data Center networks with VXLAN BGP- Lukas Krattiger Check on CiscoLive App
EVPN Part II
BRKACI-2003 Cisco ACI MultiPod Design and Deployment John Weston Check on CiscoLive App

BRKARC-3222 Cisco Nexus 9000 Architecture Tim Stevenson Check on CiscoLive App

BRKDEV-1368 Effectively Understand and Leverage YANG with Hank Preston Check on CiscoLive App
NETCONF and RESTCONF for Model Driven
Programmability
BRKACI-2508 Nexus standalone fabrics migration to ACI Ricardo Trentin Check on CiscoLive App

BRKDCN-2657 Design and Implementation of DCI Networks Yves Louis Check on CiscoLive App

BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 6
Agenda

• Introduction
• Beyond STP, from Networks to Fabrics
• Scaling Brownfield and Greenfield Data Center Networks
• Data Center Interconnect Solutions
• Feature-Specific Considerations
• Programmability, Automation & Orchestration
Some European Statistics
Out of 21M enterprises

Midsize enterprise:
Micro Enterprises
• 50 < employees < 250 Small Enterprises
Medium Enterpries
• € 10M < Revenue < € 50M Large Enterprises

http://ec.europa.eu/index_en.htm

Depends on Geography
BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 8
Designing Data Center Networks for Midsize Enterprises
Defining “Midsize”
Client Access
 Require dedicated DC Switches, moving WAN / DCI
away from collapsed core with campus
Campus

 Mostly virtualized, some physical servers


Scalability
L3
-----------
 Size for current needs, reuse components L2
in larger designs FC

Design Options FCoE


iSCSI / NAS

 Feature choice + priority = tradeoffs


 No such a things as the perfect recipe’
 Where the industry is going:
 Programmability, Automation (SDN buzz)
BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 9
What are you ready for?

Decisions will depend on where you draw the line:


• Want to stay with existing toolsets for config & management?
• Capable of consuming a new set of tools in the hope of some efficiency?
• New or traditional operational model?
• Resist/accept pressure from other teams or business owners?

BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 10
Design Goals

Flexible

Reliable

Easy

BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 11
Network Topologies

Ring Star Fully Connected Bus

Tree N-Tiered Spine Leaf

BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 12
Which network model would YOU choose?
Example: 350 VM’s; 25 Baremetal Servers @ 1/10GE; HA; IP Based Storage

VXLAN

L3
L3
… …
96 front L2 96 front L2

facing ports facing ports

Collapsed/Single Tier Spine Leaf


1RU

Downlink ports: 48p Uplink/Breakout ports: 6-12p

BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 13
Single-Tier, Dual-Tier, Spine/Leaf
Scalable Spine/Leaf DC Fabric

VXLAN

Dual Tier DC

Scale by Tiering

VXLAN

Single Layer DC

Scale Out
Small Spine/Leaf

BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 14
Connectivity & Features Drive Design Choices
 Computing Form Factor  Connectivity Model
– Unified Computing Fabric – 25 or 10 or 1-GigE Server ports
– 3rd Party Blade Servers – NIC/HBA Interfaces per-server
– Rack Servers (Non-UCS Managed) – Copper vs Fiber cabling to ToR

 Storage & Storage Protocols  Virtual Networking Requirements


– Native Fibre Channel – vSwitch/DVS/OVS/Nexus1Kv/AVS
– Unified Ports, FCoE
 Programmabiltiy/Automation/Orchestration
– IP-based storage (iSCSI, NAS)
– Complete abstraction
– Device or fabric level programmability

iSCSI NFS/
FCoE FC CIFS

VM VM VM VM VM VM

BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 15
Data Center Network Needs
• Server virtualization is here
• Higher P2V ratio, access port bandwidth up
Internet
• Oversubscription ratio from 20:1 to 1:1 Public Offsite DC
Cloud Enterprise
• Horizontal workload mobility Site B

NORTH - SOUTH TRAFFIC


Mobile Network
• Any VLAN everywhere, no physical boundaries
FC

• Highly distributed applications DATA FCoE


iSCSI / NAS
API CENTER
• Hadoop and microservices NETWORK Storage
• More server to server communication Orchestration/
Monitoring
• 80/20 traffic pattern reversed?
• Increasing high availability requirements Server/Compute Services

• Where did my maintenance window go? EAST – WEST TRAFFIC

• More execution speed, less human errors


BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 16
From Networks To Fabrics
Definition: ensemble of switches that behave
and get configured like a single giant switch
Internet Offsite DC
Public
• Flexibility: allows workload mobility, VLANs everywhere Cloud Enterprise
Site B

NORTH - SOUTH TRAFFIC


Mobile Network
• Robustness: reduce L2 failure domains, L2/L3
boundary on leafs, anycast gateway FC

DATA FCoE
• (Virtual) Network services: moving out to border API CENTER
iSCSI / NAS

leafs, policy-based service chaining FABRIC Storage


Orchestration/
• Performance: full cross sectional bandwidth Monitoring
(any-to-any) with ECMP, avoid oversubscription
• Latency: deterministic at scale, single hop away Server/Compute Services

• Scalability: add end nodes, maintain oversubscription EAST – WEST TRAFFIC

• Cost: fixed switches vs modular switches

BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 17
Network Implications of NVM/NVMe/NVMe-oF
NVM storage as a key force driving the transition to higher bit rates

• Rule of thumb: 1M IOPS need 40G of bandwidth


• 4x 10Gbps (or equivalent FC bandwidth) connectivity required in 2018
• 18x 40Gbps (or equivalent FC bandwidth) in 2022
• Making leaf-spine networks a necessity? More high-speed ports on access switches?

BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 18
Agenda

• Introduction
• Beyond STP, from Networks to Fabrics
• Scaling Brownfield and Greenfield Data Center Networks
• Data Center Interconnect Solutions
• Feature-Specific Considerations
• Programmability, Automation & Orchestration
History Lesson: Spanning tree

• Spanning Tree introduced around 1985


• 32 years ago, we also saw:
• Windows 1.0
• DNS come out of academia
• First Nintendo Entertainment System
• STP prevents loops, but blocked ports
Host or
and slow re-convergence Switch

• Successfully deployed for some time…


• …but since a few years, most DC
Designs built to work around STP

BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 20
Virtual Port Channel (VPC)

• VPC invented to overcome STP limitations vPC Domain

• IEEE standard in 2000 (802.3ad)


• Not perfect, but a good workaround
• STP is still there on every link
• Human error, misconfiguration, bug can Host or
still cause issues Switch

• Cisco alternatives for campus design:


VSS, StackWise Virtual

BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 21
Virtual Port Channel (VPC) “Mini-Fabric”
vPC Domain 1
• VPC Northbound & Southbound
• More efficient than native STP
Dual sided vPC
• STP is still running Back-to-Back vPC
“mini-fabric”
• Another good workaround
• Configuration can become complex
as switch counts grow vPC Domain 2

• vPC makes two switches look as


one….but what about 4 switches?
Host or
Switch

BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 22
Cisco Live Session: BRKDCT-2949
Building Data Center networks with VXLAN BGP-EVPN Part I

Cisco Live Session: BRKDCT-3378


L3-Based Fabrics Building Data Center networks with VXLAN BGP-EVPN Part II

• Every link forwarding


• L3 ”routing” fast convergence
• VXLAN overlay (MAC in UDP) VXLAN

• Flood and learn vs VXLAN BGP-


EVPN control plane
• STP might still exist on edges, but
not within the “fabric”
• VPC still needed at edge
Host or
• Spine/Leaf: Switch
• Flexible and efficient design
• Consistent hop count & latency
BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 23
Why Spine-Leaf Design? Flexibility and Efficiency
Scale ports and bandwidth independently

BW Fabric
40G fabric ports
M fabric links per leaf

BW Edge N ports

Oversubscription at leafs if:


10G host ports
M x BW Fabric < N x BW Edge
96 ports
2x48 10G (960 Gbps total)

BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 24
Why Spine-Leaf Design? Flexibility and Efficiency
Scale ports and bandwidth independently
To speed up FCT,
Need even more
boost the backplane
host ports?

Per Spine
Utilization
by spreading load
Add another leaf
across more spines

BW Fabric
FCT = Flow
40G fabric ports Completion Time

FCT

FCT

FCT
M fabric links per leaf

more ports
BW Edge N ports

Oversubscription at leafs if:


10G host ports
M x BW Fabric < N x BW Edge
192 ports
4x48 10G (1920 Gbps total)

BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 24
Why Spine-Leaf Design? Flexibility and Efficiency
Scale ports and bandwidth independently
To speed up FCT,
Need even more

Utilization
Per Spine
boost the backplane
host ports?
by spreading load
Add another leaf more
bandwidth
across more spines

reduce failure domain

BW Fabric
FCT = Flow
40G fabric ports Completion Time
FCT

FCT

FCT
M fabric links per leaf

more ports
BW Edge N ports

Oversubscription at leafs if:


10G host ports
M x BW Fabric < N x BW Edge
Lower FCT = FASTER
192 ports
APPLICATIONS
4x48 10G (1920 Gbps total)

BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 24
Spine/Leaf DC Fabric ≅ Large Non-Blocking Fabric

Host Host Host Host Host Host Host


1 2 3 4 5 6 7

BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 27
Spine/Leaf DC Fabric ≅ Large Non-Blocking Fabric

Host
4
Host
1

Host
5
Host
2

Host
6
Host
3

Host
7
Host Host Host Host Host Host Host
1 2 3 4 5 6 7

BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 28
Spine/Leaf DC Fabric ≅ Large Modular Switch

Card

Card
Line

Line

Host
4
Module
Host

Fabric
1

Card
Line

Card
Line

Host
5
Host

CLOS

Module
2

Fabric
Card
Line

Card
Fabric

Line

Host
6
Module

Card
Fabric

Line
Card
Line
Host
3

Card
Line
Card
Line

Host
7
BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 29
Impact of Link Speed – the Drive Past 10G Links

20×10Gbps 5×40Gbps 2×100Gbps


Bandwidth
Aggregate
200G

Uplinks Uplinks Uplinks


Bandwidth
Aggregate
200G

20×10Gbps 20×10Gbps 20×10Gbps


Downlinks Downlinks Downlinks

• 40 & 100Gbps fabric provide very similar performance for fabric links
• 40G provides performance, link redundancy, and low cost with BiDi
© 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public
Statistical Probabilities of Efficient Forwarding
Probability of 100% throughput ≅ 3%
20×10Gbps 5×40Gbps 2×100Gbps
Uplinks Uplinks Uplinks
1 2 20
Probability of 100% throughput ≅ 75%

1 2 3 4 5
Probability of 100% throughput ≅ 99%

11×10Gbps flows
(55% load) 1 2

BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 31
Impact of Link Speed on Flow Completion Times
20 Avg FCT: Large (10MB,∞) background flows
FCT (normalized to optimal) 18
16
Lower
FCT is 14 Series1
Better 12
Series2
10
Series3
8
6 Series4
4
2
0
30 40 50 60 70 80
Load (%) BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 32
Impact of Link Speed on Flow Completion Times
20 Avg FCT: Large (10MB,∞) background flows
FCT (normalized to optimal) 18
16
Lower
FCT is 14 Series1
Better 12
Series2
10
Series3
8
6 Series4
4
• 40/100Gbps
2 fabric links: ~ same FCT as non-blocking OQ-switch
0
30 40 50 60 70 80
• 10Gbps fabric links: FCT 40% worse than 40/100Gbps
Load (%)
Impact of Link Speed on Flow Completion Times
20 Avg FCT: Large (10MB,∞) background flows
FCT (normalized to optimal) 18
16
Flow Completion Time dependent on
Lower
FCT is 14 queuing and latency
Series1
Better 12
Series2
10 40G is not just about faster ports &
8 higher bandwidth, it’s about Series3

6 shorter Flow Completion Time Series4


4
• 40/100Gbps
2 fabric links: ~ same FCT as non-blocking OQ-switch
0
30 40 50 60 70 80
• 10Gbps fabric links: FCT 40% worse than 40/100Gbps
Load (%)
40G BiDi Optics Preserve Existing MM 10G Cabling
MM Fiber Plant
MMF LC
Patch cord
Used Fiber Pair
MMF LC
Patch cord
Highlights:
SFP-10G-SR SFP-10G-SR • Reuse of existing
cable plant
QSFP-40G-SR4 QSFP-40G-SR4
MM Fiber Plant • Price of 40G BiDi
Used Fiber Pair
QSFP ~ price for 10G

MPO
MPO

Used Fiber Pair


Used Fiber Pair
Used Fiber Pair
SR SFP

• WDM technology
QSFP-40G-SR-BD QSFP-40G-SR-BD over MM fibers
MM Fiber Plant
MMF LC Used Fiber Pair
MMF LC • Excellent DC reach
Patch cord Patch cord

• Can we do the same


at 100G?
Distance up to 125m with OM4 BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 35
Networking Underlays and Overlays

IP Fabric

Robust Underlay/Fabric Flexible Overlay Virtual Network


• High Capacity Resilient Fabric • Mobility – Track end-point attach at edges

• Intelligent Packet Handling • Scale – Reduce core state, distribute to edge

• Full Cross Sectional Bandwidth • Flexibility/Programmability – Less touch points

• Programmable & Manageable • Physical + Virtual; L2 & L3 Connectivity

Do you really want to deal with two networks? Consider Cisco ACI instead
BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 36
Why VXLAN Overlays?

VXLAN provides a Network with Segmentation, IP Mobility,


and Scale
• “Standards” based Overlay (RFC 7348)
• Leverages Layer-3 ECMP underlay – all links forwarding
• Increased Name-Space to 16M identifier
• Integration of Physical and Virtual Classical Ethernet Frame DMAC SMAC 802.1Q Etype Payload CRC

50 bytes Original CE Frame

• It’s the pillar of SDN Outer


IP UDP VxLAN 802.1Q CRC
MAC DMAC SMAC Etype Payload
(20) (8) (8) optional (new)
(14)

• Works better with a control plane 8 bits 24 bits 24 bits 8 bits

ags Reserved VNIVNI Reserved

BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 37
Network Centric Deployment
• Leverage known networking constructs.
• VLANs, IP address / subnets, Flood domains etc.
Existing network – Nexus, Catalyst, etc
VRF VNI: 300100 VRF VNI: 300100

MAC_A,IP_A: 10.1.1.2 MAC_B,IP_B:20.1.1.2


VNI L2: 100100 VNI L2: 100200
VLAN 100:
10.1.1.0/24 VLAN 100 VLAN 200

VLAN 200:
20.1.1.0/24 VXLAN EVPN Fabric

Existing network – Nexus, Catalyst, etc TENANT

PRIVATE NETWORK (VRF)

BD: Blue BD: Red


10.1.1.0/24 20.1.1.0/24
VLAN 100:
10.1.1.0/24 EPG: Blue-100 EPG: Red-200
VLAN 200:
20.1.1.0/24 ACI Fabric

BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 38
Cisco Live Session: PSOACI-4592
Why ACI: an overview and a customer (BBVA) perspective

Application Centric Deployment (ACI only)

• Leverage grouped constructs – Network Level; Application Level


• Abstraction with Application Network Profiles, Policy oriented
• Automation of network services
• APIC Controller: declaration of intention, translate to network constructs via policy

External
Connectivity L3 Outside

APIC

EPG EPG EPG


Web App DB
Web App DB

FW / SLB FW / SLB
App Profile

BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 39
Controlling, Managing & Automating the Fabric
Application Policy
Datacenter Infrastructure
Network Manager Client Access Controller Client Access
Connection

Creation Expansion
WAN / DCI WAN / DCI

Reporting Fault Mgmt


APIC APIC
APIC

Network Centric only Network Centric


Policy Centric

Spine Spine

VXLAN EVPN ACI Fabric

Leaf Leaf

Programmable Fabric Application Centric Infrastructure (ACI)


Variety of Nexus models Nexus9000 only

BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 40
ACI Anywhere
Any Workload, Any Location, Any Cloud
ACI ANYWHERE

Remote PoD Multi-Pod / Multi-Site Hybrid Cloud Extension

IP IP
WAN WAN

Remote Location On Premise Public Cloud

Security Everywhere Analytics Everywhere Policy Everywhere

BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 41
Agenda

• Introduction
• Beyond STP, from Networks to Fabrics
• Scaling Brownfield and Greenfield Data Center Networks
• Data Center Interconnect Solutions
• Feature-Specific Considerations
• Programmability, Automation & Orchestration
Scaling a VPC-based DC design

Access L3
Layer L2
VLANs
100-150 Host Host Host

BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 43
Scaling a VPC-based DC design

Consolidated
Core/Agg
Layer

Access L3 Access
Layer L2 Layer
VLANs VLANs
100-150 Host Host Host Host Host Host 151-200
BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 44
Scaling a VPC-based DC design

Distributed or
Consolidated
Core/Agg
Layer
L3
L2

Access Access
Layer Layer
VLANs VLANs
100-150 Host Host Host Host Host Host 151-200
BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 45
Cisco Live Session: BRKACI-2508
Integrating ACI with an existing network Nexus standalone fabrics migration to ACI

http://www.cisco.com/c/en/us/td/docs/switches/datacenter/aci/apic/sw/migration_guides/migrating_existing_networks_to_aci.html

BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 46
Integrating ACI with an existing network
ACI Pod
New DC
Row Upgrade
New Application
Distributed or
Consolidated
Core/Agg Spine
Layer Layer
L3
ACI Fabric
L2 (VXLAN based)
Dual sided
vPC

ACI Border Leafs


Access Access Access
Layer Layer Layer
VLANs VLANs VLAN
100-150 151-200
Host 201-250
Host Host

BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 47
Integrating ACI with an existing network
ACI Pod
New DC
Row Upgrade
New Application
Distributed or
Consolidated
Core/Agg Spine
Layer Layer
L3
ACI Fabric
L2 (VXLAN based)

Access ACI
Layer
VLANs
Leafs
100-150
Host and Border Leafs
Host Host

BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 48
Supported from ACI
Remote Leaf Concept 3.1(1) – Dec 2017

Remote Leaves
Any router with OSPF, logically connect
APIC and Spine Nodes (Proxy) DHCP relay and 9150
MTU support.
to Spines of one
remain at primary site(s) IP WAN POD in Main DC
IPN L2 / L3

1/10G ports
Tested with 500Mbps on WAN,
but the more the better

All local traffic is switched directly between


endpoints, both virtual and bare metal
Specific traffic requires use of spine proxy.
vSwitch
Hypervisor
Supported hardware and scale
Remote site contains Nexus 9300 connected
- All hardware from -EX onwards is supported
to IP Network via uplink ports and fully
- 40 RL (20 sites, 2 RL per site)
managed by primary site APIC cluster

BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 49
ACI: Small Footprint Option Under Consideration

• Full featured ACI Fabric


Physical Server
• Reduced initial port count
Spine 32 ports

Spine 32 ports
• Optimized scalability for
Leaf 48 ports VM mid size deployments
Leaf 48 ports VM
• Lower entry cost (-30%)

What is your
opinion? • Reduced footprint (-40%)
BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 50
Agenda

• Introduction
• Beyond STP, from Networks to Fabrics
• Scaling Brownfield and Greenfield Data Center Networks
• Data Center Interconnect Solutions
• Feature-Specific Considerations
• Programmability, Automation & Orchestration
Cisco Live Session: BRKDCN-2657
Design and Implementation of DCI Networks
IT Trends - Distributed Data Centers
Building the Data Center Cloud
Distributed Data Center Goals:
• Seamless workload mobility between multiple
datacenters
• Distributed applications closer to end users
• Pool and maximize global compute resources
• Ensure business continuity and disaster
avoidance with workload mobility, distributed
deployments and clustered applications
• Midsized enterprises often rely on two Geographically
datacenters Dispersed Data Centers

BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 52
NX-OS Overlay Transport Virtualization Technology
Extend VLANs Across Datacenters

OTV

Classical Pod Scalable Pod Leaf Spine Pod


Spanning Tree Protocol vPC, N-tier design VXLAN, ACI

Feature: Benefits: In the news:


• Simplified Layer 2 / VLAN Extensions • Field proven, very mature • M3 linecards with NX OS 8.0
• Restricted fault domain (STP domain), • Enable L2 elasticity across DC’s • OTV Loopback Join Interface for
loop prevention Multicast-based OTV control-plane
• Simplify Virtual Machine Mobility
• Optimized Multicast replication • Wire-rate 256-bit AES MACsec on all
• Extend Layer 2 without the risks of
ports at all speeds along with OTV
• Dual homing large fault domain
• Works over dark fibre, MPLS or IP • Simple – 3 easy commands !

BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 53
Client Access

OTV options for DCI WAN / DCI


Client Access Campus

WAN / DCI

Campus

L3
-----------
L2
L3
-----------
L2

Virtual DC Virtual DC
Services in Services in
Software VM VM VM VM VM VM VM VM VM VM VM VM Software

Virtualized Servers, Nexus CSR1000v Virtualized Servers, Nexus


1000v, vPath, CSR 1000v 1000v, vPath, CSR 1000v

BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 54
Client Access

OTV options for DCI ASR1000


WAN / DCI
Client Access Campus

WAN / DCI

Campus

ASR1000 L3
-----------
L2
L3
-----------
L2 N7K

Physical or Physical or
Virtual Virtual
Workloads Workloads
VM VM VM VM VM VM VM VM
and Services and Services

CSR1000v

BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 55
VXLAN as a Data Center Interconnect? Building

• DCI is an architectural discussion

• VXLAN is just an encapsulation technique

• OTV is a well-proven technology that gets you


fully covered

• VXLAN can absolutely fit into a DCI


architecture…if you handle it CAREFULLY
• Protection at edge as well as control plane required
• Storm Control, BPDU Guard, HMM Route Tracking
• Control-Plane with MAC-learning, ARP suppression

© 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public
Cisco Live Session: BRKACI-2003
ACI MultiPOD Design and Deployment

Cisco Live Session: BRKACI-2125


DCI with ACI Multi-Fabric Options ACI Multi-Site Architecture and Deployment

Single APIC Cluster/Single Domain (BC-DA) Multiple APIC Clusters/Multiple Domains (DR)
Stretched Fabric Multi-Fabric Connected (L2 and L3 Extension)
ACI Fabric ACI Fabric 1 ACI Fabric N
Site 1 Site 2

40G
L2/L3

More scale
Multi-Pod More flexibility Multi-Site (new)
More automation
Pod ‘A’ Inter Pod Pod ‘n’
Network@10G Site ‘A’ Inter Pod Site ‘n’
Network@10G
MP-BGP - EVPN

… MP-BGP - EVPN

APIC Cluster
Agenda

• Introduction
• Beyond STP, from Networks to Fabrics
• Scaling Brownfield and Greenfield Data Center Networks
• Data Center Interconnect Solutions
• Feature-Specific Considerations
• Programmability, Automation & Orchestration
What Features Matter The Most to You? (as of Jan 2018)
Feature 3000 3500 5500 5600/ 7000/ 9300 9500
6000 7700
Unified Ports

FCoE

FEX

VXLAN Bridging and Routing

VXLAN F&L

VXLAN BGP EVPN

DCI - OTV

DCI – VXLAN

ISSU

VDC

LISP, MPLS

ITD

RISE

Extreme low latency

IPv6 support

Automation ACI

Automation DCNM

BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 59
FCoE and Unified Fabric
Fibre Channel with simpler infrastructure and lower cost

FCoE Benefits

Encapsulate Fibre Channel frames


into Lossless Ethernet • Standards based
• Operationally same as existing LAN /SAN
Ethernet Carrier, DCB
• Transparent to OS and Apps
Individual • Fewer Cables
Ethernets
• Fewer switches
Individual • Fewer adapters
Storage
(IP, Eth, FC) • Overall less power

Byte 0 Byte 2229


Ethernet
Header

Header

Header
FCoE

CRC
EOF
FCS

FC Payload
FC

BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 60
Nexus Fabric Switches Evolution (FC/FCoE View)

BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 61
Nexus 9300 Unified Ports
Compelling Business Value and Flexibility

 Repurposing of any switch port as either


CAPEX Fibre Channel or Ethernet/FCoE avoids
CAPEX lock-in with specific port-types

Unified
Port
 Reduces physical management points
OPEX
 Extend adoption of FCoE in the
access with associated savings FC Eth

Native Fibre Lossless


 Simplifies the purchasing decision, makes
BUSINESS port-type decisions a thing of the past Channel Ethernet:
AGILITY 16/32G 1/10/25GbE, FCoE,
 Allows IT to respond faster to changing iSCSI, NFS/SMB
business needs

BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 62
Introducing 16/32G FC on Nexus 93180YC-FX

Milestone Timeline Shipping


FCoE NPV mode Oct, 2017
(NX-OS and ACI) NX OS 7.0(3)I7(1)

16G Fibre Channel Nov, 2017


NPV mode (NX-OS) NX OS 7.0(3)I7(2)

32G Fibre Channel Soon


NPV mode (NX-OS)
Unified Ports: 48p 1/10/25G or 16/32G FC SFP 6p 40/100G QSFP
16/32G Fibre Channel Work in progress
NPV mode (ACI)

First Nexus 9300 switch with Fibre Channel Capabilities


Supports NX-OS and ACI Mode.

Seamless transition Multi-protocol UNIFIED PORT Interop with Cisco Cloud Scale
from N5K FC MDS and Nexus 5K technologies
FCoE
IP Storage MACSec

EX only supports FCoE NPV


FX supports both FCoE NPV (Oct 2017) and FC NPV (Nov 2017)
BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 63
Example: Small Deployment For All Flash Array
All Flash Array

NPIV
feature
32G FC storage ports

MDS 9132T
16G FC uplink ports
MDS 9132T LAN
NPV mode (32G future)

Nexus 93180YC-FX Peer links (Ethernet) Nexus 93180YC-FX

10G FCoE access ports

NX OS mode: shipping
ACI mode: work in progress

Servers

BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 64
What is ITD ?
Intelligent Traffic Director

• Traffic distribution and redirection

• ASIC based solution (HW-switched)

• Caters to multi-terabit traffic

• Works on Nexus switches – 5/6/7/9k,


similar but not identical capabilities
ITD does L3-L4 traffic distribution,
does not replace Layer-7 Load-balancers

BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 65
Where to use ITD ?

ITD to load-balance to the destination


#1
Example: Server-Load Balancing

Clients Servers
Health checking supported
BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 66
Where to use ITD ?
ITD for In-line traffic redirection
#2 Example: Firewalls, IDS, Web Caches etc.
Possible WCCP replacement? Possible MS NLB replacement?

Destination

Clients
Firewalls/other appliances

BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 67
ITD Benefits
Intelligent Traffic Director
No service-module
Line-rate
or external
traffic-distribution
appliance required

Automatic Ease of deployment,


failure handling reduced configuration

BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 68
Agenda

• Introduction
• Beyond STP, from Networks to Fabrics
• Scaling Brownfield and Greenfield Data Center Networks
• Data Center Interconnect Solutions
• Feature-Specific Considerations
• Programmability, Automation & Orchestration
Why Network Programmability?

More Speed More Repeatability More Flexibility More Innovation

Fewer Mistakes

BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 70
Programmability, Automation, Orchestration

BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 71
Cisco Data Center Network Automation Types
Providing Choice in Automation and Programmability
Application Centric
Programmable Network Programmable Fabric
Infrastructure

DCNM

DB DB

Web Web App Web App

Modern NX-OS with enhanced VxLAN-BGP EVPN Turnkey integrated solution with
NX-APIs standard-based security, centralized management,
compliance and scale
DevOps toolset used for Network 3rd party controller support
Management Automated application centric-policy
(Puppet, Chef, Ansible etc.) Cisco Controller for software model with embedded security
overlay provisioning and
management across N2K-N9K Broad and deep ecosystem

Tools, APIs, Controllers and Automation© 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public
Cisco Live Session: BRKDEV-1368
Effectively Understand and Leverage YANG with NETCONF and
RESTCONF for Model Driven Programmability

Nexus Programmability Nexus 7K Nexus 5K / 6K Nexus 9K


Puppet/Chef/Ansible Shipping Shipping Shipping
Provisioning &
PoAP Shipping Shipping Shipping
Orchestration
OpenStack Shipping Shipping Shipping

NetConf (SSH/XML) Shipping Shipping Shipping


NX API CLI (JSON/XML) Shipping Shipping Shipping

Protocols and NX API REST Future No Shipping


Data Models NetConf / YANG Future No Shipping
RESTconf/gRPC Future No Shipping
Streaming Telemetry Future No Shipping

Native Python Shipping Shipping Shipping


Linux container Shipping Shipping Shipping
Programmatic
Docker containers Future No Future
Interfaces
Guest Shell Future No Shipping
OpenFlow Shipping No Shipping

BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 73
Programming a Fabric
• A lot of work is being done to provide customers maximum flexibility in
programming & automation interfaces
• Free Open Programmability book:
• http://www.cisco.com/c/dam/en/us/td/docs/switches/datacenter/nexus9000/sw/open_nxos/program
mability/guide/Programmability_Open_NX-OS.pdf
• New community site dedicated to NXOS programmability:
• https://opennxos.cisco.com

• A lot of work has been done to increase available knowledge on network


programming across all Cisco products
• DevNet: If you haven’t visited, please do so
• https://devnet.cisco.com
• SANDBOX! – FREE 24 X 7 hosted labs

BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 74
Cisco Data Center Network Automation Types
Providing Choice in Automation and Programmability
Programmable Application Centric
Programmable Fabric
Networking Infrastructure

DCNM

DB DB

Web Web App Web App

Modern NX-OS with enhanced VxLAN-BGP EVPN Turnkey integrated solution with
NX-APIs standard-based security, centralized management,
compliance and scale
DevOps toolset used for Network 3rd party controller support
Management Automated application centric-policy
(Puppet, Chef, Ansible etc.) Cisco Controller for software model with embedded security
overlay provisioning and
management across N2K-N9K Broad and deep ecosystem

Tools, APIs, Controllers and Automation© 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public
Along the Spectrum from CLI to ACI
A New Way To Do Fabric Management

Scripting to the
CLI and/or API

Basic
Element
Manager

CLI ACI

BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 76
Fabric Management Functionalities
Common Customer Asks
• Fabric management automation – high interest
• Many Fabrics are based on things like VXLAN, BGP/EVPN, IS-IS
• New protocols, new configurations, new things to learn
• Simple tool to ease burden of adoption
• Simplified interaction – GUI and API
• CVD/Best practices – done for you!

BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 77
Data Center Network Manager (DCNM) 10

• Network management system tool for


• SAN management (Nexus and MDS)
• LAN (full Nexus family)
• Programmable Fabrics (NX OS mode)
• Flow controller for IP Media Networks
• Advanced feature support (STP, VXLAN, POAP,
OTV, templates, topology, configuration backup)
• Underlay/Overlay manager for VXLAN-EVPN (IP)
Fabrics in NX-OS Mode

BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 78
Functions Dashboard

Health Monitoring, Configuration/Backup Automation Trend Analysis & VM


Inventory & Analytics
Diagnostics

Host / Endpoint Visualization & Alert/Notifications Storage Management Professional Media


Monitoring Troubleshooting Net Controller

BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 79
Cisco Data Center Network Automation Types
Providing Choice in Automation and Programmability
Programmable Application Centric
Programmable Fabric
Networking Infrastructure

DCNM

DB DB

Web Web App Web App

Modern NX-OS with enhanced VxLAN-BGP EVPN Turnkey integrated solution with
NX-APIs standard-based security, centralized management,
compliance and scale
DevOps toolset used for Network 3rd party controller support
Management Automated application centric-policy
(Puppet, Chef, Ansible etc.) Cisco Controller for software model with embedded security
overlay provisioning and
management across N2K-N9K Broad and deep ecosystem

Tools, APIs, Controllers and Automation© 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public
FCAPS and Automation
Application Centric
Programmable Network Programmable Fabric
Infrastructure

DCNM

Turnkey integrated solution with security, centralised


Modern NX-OS with enhanced NX-APIs Integrated stack management, compliance and scale
Or
DevOps toolset used for Network Management A-la-carte Automation Automated application centric-policy model with
(Puppet, Chef, Ansible etc.) embedded security
Streamlined Workflow Management
Custom Script based Operations and Workflows
Broad and deep ecosystem

Security
External
Performance Tools
External
Accounting Tools Integrated
Integrated Tools
Fault
Tools
Configuration

BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 81
Effort vs Time – Traditional IT Build/Run

Ops
Config
Ops

Config

Ops
Config

Config Ops

BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 82
Effort vs Time – Traditional IT Build/Run

Ops
Config
Ops

Config

Ops
Config

Config Ops

BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 83
Effort vs Time – Utilizing Abstraction & Automation
Plan Implement
Ops
Config
Ops

Config
Ops

Ops
Config

Config Ops

BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 84
Cisco UCS Director for Compute; Network & Storage
Infrastructure consumption made easier
Converged Stack
Secure Cloud Control Panel
Container
OS and
Network Compute VMs Storage Virtual Bare
VM VM
Metal
Machines
Single Pane of Glass for Virtualized and Bare-Metal

Policy-Driven Virtual AND Physical


Compute
Provisioning Speed with Accuracy Compute and Hypervisor

UCS Director More efficient use of People


Network A B C
& Time
Network and Services

Consistency, Less Error in


Tenant Tenant Tenant
Repetitive Tasks Storage A B C

BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 85
UCS Director – Example Use Cases for SMB

 Provision new SP from template, provision zoning/masking/boot LUN, boot from SAN,
install hypervisor (ESX image on PXE server), associate VMWare Host Profile for vSwitch
settings, more customization
---------
-
------------
---------
------------
Start ------------
------------
-
---------
---------
------
---------
---

 Create VM Disk, install and format in OS, then create windows share or linux mount point

Start

BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 86
Key Takeaways
 Cisco has many options for building data
center networks, you own the key
 All solutions can start small and grow
 No Cisco solution has to be a “rip and replace”
 Spine-Leaf does not have to be expensive
 Programmable fabrics provide new tools for
simplified operations
 Automated fabrics provide new methods of
managing DC Networking

BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 87
Call to Action

 Check out the sessions listed at beginning of presentation


at www.ciscolive.com
 Visit the World of Solutions and look for “Data Center”
 Meet the Engineer
 Visit the DevNet Zone and see what they have to offer!

BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 88
Cisco Spark
Questions?
Use Cisco Spark to communicate
with the speaker after the session

How
1. Find this session in the Cisco Live Mobile App
2. Click “Join the Discussion”
3. Install Spark or go directly to the space
4. Enter messages/questions in the space

cs.co/ciscolivebot#BRKDCT-2218

© 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public
• Please complete your Online Complete Your Online
Session Evaluations after each
session
Session Evaluation
• Complete 4 Session Evaluations
& the Overall Conference
Evaluation (available from
Thursday) to receive your Cisco
Live T-shirt
• All surveys can be completed via
the Cisco Live Mobile App or the
Communication Stations
Don’t forget: Cisco Live sessions will be available
for viewing on-demand after the event at
www.ciscolive.com/global/on-demand-library/.

© 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public
Continue Your Education
• Demos in the Cisco campus
• Walk-in Self-Paced Labs
• Tech Circle
• Meet the Engineer 1:1 meetings
• Related sessions

BRKDCT-2218 © 2018 Cisco and/or its affiliates. All rights reserved. Cisco Public 91
Thank you

You might also like