Aci tech talk
-
Upload
cisco-public-sector -
Category
Technology
-
view
399 -
download
0
Transcript of Aci tech talk
2 © 2015 Cisco and/or its affiliates. All rights reserved. Cisco Confidential
• Evolution of DC Switching Design • Application Centric Infrastructure
Session Agenda
4 © 2015 Cisco and/or its affiliates. All rights reserved. Cisco Confidential
How Cisco Data Center Switching Has Evolved From the Catalyst 6500 to the existing Nexus 9000 family
Catalyst 6000 / 6500 Flagship Campus and Datacenter switch with 100/1GE market leadership
Nexus 7000 1st product on Nexus family
aimed to transition from Cat6K with 1/10GE focus
Nexus 2K / 5K / 6K Launched as part of the UCS foundation (FEX and FCoE as innovation). Transition to ToR-based designs
Nexus 3K / 3500 Nexus 3064 was the 1st DC
switch with merchant silicon (targeting HFT initially)
Nexus 3548 followed with Cisco ASIC
2000 /
2003 2007 From
2008 2011
Nexus 7700 N7K architecture evolution of bringing higher performance and 100GE
2012 2014
Nexus 9K / ACI Foundation to ACI with focus on 10GE
and 40GE at scale with options for
25/50/100GE
5 © 2015 Cisco and/or its affiliates. All rights reserved. Cisco Confidential
Networking Today
Designed for North / South campus traffic
East / West traffic is inefficient & can be unpredictable
Bandwidth & latency is not deterministic
6 © 2015 Cisco and/or its affiliates. All rights reserved. Cisco Confidential
How Cisco Data Center Switching Design Has Evolved Fabric Designs - Escaping Spanning Tree
VSS/L3 Access
vPC
FEX
Fabricpath
VXLAN/ACI/SDN
VLANs
RPVST
Workload Mobility/vMotion
OTV
Multi-site/DR Hybrid Cloud Security L4-L7 Services
TRILL MPLS
Discrete L2 Networks Integrated L3 Fabric
East-West Traffic
7 © 2015 Cisco and/or its affiliates. All rights reserved. Cisco Confidential
Extending the Switch Fabric
Switch Fabric on Supervisor
Fabric Extender - Nexus 5k/2k Catalyst 6500
10GE Fabric Extension
8 © 2015 Cisco and/or its affiliates. All rights reserved. Cisco Confidential
Distributing the Switch Fabric
Spine Layer ~ Fabric Modules
Leaf Layer ~ Line Cards
40G/100G per Spine Connection 10G/25G/40G per leaf connection
APIC (in ACI) ~ Supervisor
Nexus 7706 Spine-Leaf Fabric
Supervisor: Control Plane/Configuration
Line Cards: Host-facing ports
Fabric Modules: Connectivity between line cards. System bandwidth can be increased by adding or upgrading modules.
SDN technologies such as VXLAN overlays and SDN controllers make distributing the switching components possible. An L3 Spine-Leaf switch fabric eliminates spanning-tree and builds a large virtual L2/L3 switch optimized for East-West traffic flows.
VXLAN
9 © 2015 Cisco and/or its affiliates. All rights reserved. Cisco Confidential
APIC Controller
Overview of the ACI Fabric
ACI Spine Nodes
ACI Leaf Nodes
ACI Fabric Features - ACI Spine Layer – Provides bandwidth and redundancy between Leaf Nodes
ACI Leaf Layer – Provides all connectivity outside the fabric - including servers, service devices, other networks
Optimized Traffic Flows – Accommodates new E-W traffic patterns in simple, scalable, non-blocking design
Decoupling of Endpoint Identity – Network policies automatically move with VM/Server/Container
Network Innovations – Dynamic load balancing, dynamic packet prioritization, congestion management
ACI Fabric
Industry’s most efficient fabric: - 220k+ 1/10/25/40Gb edge hosts - High-density 40/100G spine - 1 million+ IPv4 / IPv6 endpoints - 64,000+ tenants
10 © 2015 Cisco and/or its affiliates. All rights reserved. Cisco Confidential
What problem are we solving?
11 © 2015 Cisco and/or its affiliates. All rights reserved. Cisco Confidential
Now let’s imagine a network switch … … at the moment, largely configured on the CLI
12 © 2015 Cisco and/or its affiliates. All rights reserved. Cisco Confidential
• Device basics: AAA, syslog, SNMP, PoAP, hash seed, default routing protocol bandwidth …
• Interface and/or Interface Pairs: UDLD, BFD, MTU, interface route metric, channel hashing, Queuing, LACP, …
• Fabric and hardware specific design: HW Tables, TCAM, …
• Switch Pair/Group: HSRP/VRRP, VLANs, vPC, STP, HSRP sync with vPC, Routing peering, Routing Policies, …
• Application specific: ACL, PBR, static routes, QoS, ...
• Fabric wide: MST, VRF, VLAN, queuing, CAM/MAC & ARP timers, COPP, route protocol defaults
All nodes are managed and operated independently, and the actual topology dictates a lot of configuration
13 © 2015 Cisco and/or its affiliates. All rights reserved. Cisco Confidential
Cisco ACI solves the problem … Interfaces, protocols, TCAM, etc … all represented in an object model, and ALL accessible through an XML/JSON API and CLI (yes, you can still use the cli)
14 © 2015 Cisco and/or its affiliates. All rights reserved. Cisco Confidential
APIC becomes single point of management for the entire fabric…with a policy-based model
15 © 2015 Cisco and/or its affiliates. All rights reserved. Cisco Confidential
…and the fabric becomes the Data Center switch
16 © 2015 Cisco and/or its affiliates. All rights reserved. Cisco Confidential
Adding, removing or replacing nodes becomes extremely simple
17 © 2015 Cisco and/or its affiliates. All rights reserved. Cisco Confidential
And so do network upgrades …
18 © 2015 Cisco and/or its affiliates. All rights reserved. Cisco Confidential
… and you get best troubleshooting with full physical, virtual and services visibility …
19 © 2015 Cisco and/or its affiliates. All rights reserved. Cisco Confidential
System Health Scores
Statistics Per App
Endpoint Troubleshooting
Wizard
Contract Deny Logs
Real-time Heat Maps
Endpoint Tracker
ACI – Day 2 Tools for Simplified Operations
20 © 2015 Cisco and/or its affiliates. All rights reserved. Cisco Confidential
So, the first thing to remember about ACI: it is a programmable fabric with a single point of management.