The final day of videos!!, time to get this done and start with the review of data for the test.
Section 38 continued…
- APIs (Application Programming Interfaces) – CRUD, REST and SOAP
- NETCONF and RESTCONF are APIs specifically desigened to work with network devices.
- CRUD – Create – Read – Update – Delete
- SOAP – Simple Object Access Protocol
- Statndared communicatoin protocol system taht permits processes using different OS like Linux and Windows to talk.
- typically uses HTTP(S) and data is always formatted as XML
- It has strict standard to adhere to
- REST – Representational State Transfer
- Is an architecture, not a protocol
- Gives guidelines for structure and organization of an API
- Support any transport and data format
- Commonly uses http(s) as transport and commonly uses JSON (or XML) data formats
- Typically has faster performation and is easier to use than SOAP
- Contraints
- Client- Server architecture
- Send & response model
- Uniform interface
- Statelessness
- no client context is stored on the sever between requests
- Cacheability
- responses need to be define cacheable or non-cacheable
- Layered system
- devices between client and server must be transparent
- Optional – Code on demand
- Server can temporarily extend or customze the functionality of a client by transferriing executable code.
- Client- Server architecture
- REST Response Codes
- Success
- 1xx: Informational
- 2xx: Success
- 200: OK
- 201: Created
- 204: No Content/Deleted
- 3xx: Redirection
- Errors
- 4xx
- 400: Bad request/malforned syntax
- 401: Unauthorized
- 403: Forbidden
- 404: Not Found
- 5xx: Server Error
- 400: Internal Server Error
- Responses to Get requests include data in the body
- Headers can also be included in the response
- 4xx
- Success
- Is an architecture, not a protocol
- Configuration managemetn Tools
- System to make managing large number of devices easy
- Automate many different systems from one location
- Popular free options (Enterprise paid options)
- Ansible
- Any machine with python 2 or 3
- Agentless
- Push Model
- Uses SSH by default
- Simpler than othe tools
- Modules are pre-built python scripts
- Inventory files define all hosts to be managed
- Playbooks are YAML files that outline the instructions it needs to run
- Puppet
- Uses agent on the target
- Runs on Linux
- Pull model so agent checks in every 30 minutes by default
- Programmed in Tuby
- Proprietary DSL rather than YAML
- ‘Manifest’ defines the device’s properties
- Can check config consistency
- Chef
- Uses agent on the target
- Pull model
- Programmed in Ruby
- Terminology is Cook Book > Recipe
- Ansible
- Automate provisioning an deployment of servers and network devices
- Require little program knowldege
- Development practices
- version control
- testing
- SDN – Software Defined Networking
- Router and Switch Planes
- Data (Forwarding) Plane: Traffic which is forwarded through the device
- Control Plane: Makes decisions about how to forward traffic.
- Packets such as routing protocol or spanning tree updates are destined to or locally originated on the device itself.
- Management Plane
- The device is configured and moniorted in this plane.
- SDN – Data and Control Plane Separation
- Traditionally network infrastructure devices are responsible for thier own control and data planes
- SDN decouple the data and control plans
- Network Devices are still responsible for forwarding traffic, but control plane moves to a centralised SDN controller.
- Data and control plane separation
- Rules for packet handling are sent to the network devices from the controller
- The network devices query the controller for guidance as needed and provide the controller with info about traffic being handled.
- Pure SDN vs Hybrid SDN
- Pure SDN
- The control plane runs purely on the SDN controller
- Data plane runs purely on the network device
- Hybrid SDN
- Majority of the Control plane intelligence is provided by an SND controller
- Network Devices retain some control plane intelligence as well as the data plane operations
- Most implementations use a hybrid SDN
- Pure SDN
- SDN Architecture
- Cisco SDN Controllers
- APIC – Application Policy Infrastructure Controller
- Main component of the Cisco ACI (Application Centric Infrastructure) solution
- made to manage data center environments with Nexus Switches
- DNA Center – Digital Network Architecture
- mad to manage enterprise environment
- Campus
- Branch
- WAN
- Upgrade to APIC-EM (Application Policy Infrastructure Controller – Enterprise Module
- mad to manage enterprise environment
- APIC – Application Policy Infrastructure Controller
- Router and Switch Planes
- DNA Center
- One part of three building blocks of Cisco DNA and Software Defined Architecture
- “Cisco DNA enables you to streamline operation and facilitate IT and business innovation.”
- Intent-based networking (IBN) built on Cisco DNA takes a software-delivered approach to automate and assuring services across your WAN, Campus, and Branch networks
- A Cisco SDN controller which is designed to manage enterprise environments (campus, branch and WAN)
- Opposed to APIC which manages data center environments with Nexus Switches
- DNA Center Appliance
- Runs on Cisco UCS server hardware
- Built Linux OS
- Can cluster for redundancy
- IBN – Intent Based Networking
- Controller led network that translates the business needs into policies that can be automated and applied consistently
- Goal is to continuously monitor and adjust network performance to help assure desired business
- Example 1
- a QoS policy roll-out
- The Intent: The network policy is first defined, guaranteed service to voice and video across network locations
- Traditional method
- Network team researches and plans the implementation, then configures each network device individually
- Different network device models require different commands
- Very time consuming and prone to mistakes
- IBN
- Network team creates a application policy in DNA Center specifying voice and video as business relevant apps.
- DNA Center auto configures the best practice QoS setting on devices
- Reduce time from potentially months to minutes
- a QoS policy roll-out
- Example 2
- SSecurity traffic flows in the campus
- The intent: Users in DeptA and DeptB must have connectivity to other users in the own department and the company servers
- the must not have connection to the users in the other department.
- Traditional
- Plan VLAN, IP subnet and ACL implementation, the configure each switch individually
- Users will stay in the same access port, and they are assigned a VLAN and IP address based on their physical location
- Time consuming, liable to mistakes and does not support mobility
- IBN
- Create a group-based access control policy in DNA which specifies the allowed traffic flows
- User log in from and can move to any physical location on campus
- User are Authenticated by Cisco ISE Identity Services Engine and assigned a Security Group Tag controlling their access
- SSecurity traffic flows in the campus
- Network Plug and Play
- Allows routers, switches and WAP to be deployed in remote offices with zero touch config
- Device is physically installed at location and connected to the network.
- It discovers DNA Center through various methods
- DHCP option 43
- DNS ‘pnpserver.domain-name.com’
- Device registers with and downloads its configuration from DNA Center
- This ensure correct config in remote devices and can keep a engineer to be onsite.
- Assurance
- Guarantees that the infrastructure is doing what was intended
- Receives info from all devices and ISE
- Correlation engine can identify 150+ different types of network and client issues
- Report problems and provides recommended remediation actions
- Admins can drill down into the health status of devices and clients
- See current and historical status info
- Userful for troubleshooting intermittent problems and/or issues which occurred in the past.
- Path Trace
- Use to query DNA center for the path that traffic takes over the network
- This aids troubleshooting
- API Support
- Can provide Northbout REST API
- Supports East and West bound APIs for integration for items like reporting and analytics
- Software Defined Architecture – SD Access
- Traditional Access Control
- Control access to and traffic flows within a network is fixed VLANs, IP addresses and ACLs
- User expect to connect to the same physical port and assigned VLAN and IP subnet
- ACL controls traffic flows
- More complex with each device
- Does not support mobility for the user
- SD-Access – Software Devined Access
- Newer method of network access control wich solves the limitation of the traditional implementation
- Traffic flow security is base on user identity, not physical location and IP address
- User log in from and can move to any physical location in the network.
- Tow components required
- ISE (Identity Service Engine) for user authentication
- Security Policy configured on DNA Center
- Permit and deny communication between groups
- Underlay and overlay Network
- Underlay network is the underlying physical network. Provides the underlying physical connection which the overlay network is built on top of.
- Overlay network is a logical eopolgy used to virtually connect devices. It is built over the physical underlay network.
- The combination of underlay and overlay forms the SD0Access ‘network fabric’
- Underlay Network
- Any existing (‘brownfield’) underlying physical network can be deployed.
- Links between devices can be layer 2 or layer 3 and any routing protocol can be used
- DNA center can be used to auto provision the underlay network in new (‘greenfield’) sites. Layer 3 links are used between devices and IS-IS is used as the routing protocol
- Any existing (‘brownfield’) underlying physical network can be deployed.
- Overlay Network
- LISP is used for Control Plane
- VXLAN is used for the Data Plane
- Cisco TrustSec CTS is used for the policy
- Each technology has been optimized for SD-Access
- Policy Plane – Cisco TrustSec CTS
- Users are authed by ISE
- Polciy configured on DNA Center
- Users are allocated an SGT – Scalable Group Tag
- Cisco TurstSec secures traffic flows based on the security policy and SGTs
- Standard Trust Sec needs end-to-end TrustSec devices, SD-Access uses overlay tunnels so can work with other devices.
- Traditional Access Control
- SD-WAN
- Traditional WAN Deployment
- Individual device conf
- conf is not standard
- Focus on link connectivity notthe required performance for apps
- Typically difficult to migrate to another WAN service
- It provides automated setup of WAN connectivity between sites
- Monitoring and ailover is automated
- Traffic flow control is application aware
- SD-WAN Benefits
- Auto, standardized setup of connectivity between sites
- Transport independent
- Simplified, integrated operations
- More flexibility and easier to migrate WAN services
- The required, predictable performance for important applications
- Integration with the latest cloud and network tech
- Lower Cost
- Data Plane – vEdge Routers
- vEdge router run the data plane
- They are physical or virtual routers
- They form an IPsec encrypted data plane between each other
- A site can have 2 vEdge routers for redundancy
- Control Plane – vSmart Controllers
- vSmart controllers run the control plane
- The are the centralized brain of the solution
- They run as virtual machines
- They distribute policy and forwarding info to the vEdge routers inside TLS tunnels
- Each vEdge router connects to two vSmart controllers for redundancy
- Management Plane – vManage NMS
- The vMange NMS provides the management plane GUI
- It enables centralized configuration and simplifies changes
- It provides real time alerting
- It runs as a virtual machine
- Multiple vMange NMS are clusters for redundancy
- Orchastration – vBond orchestrator
- The Bond orchestrator authenticates all vSmart controllers, vManage NMS and VEdge routers that join the SD-Wan network
- It enable vEdge routers to discover each other, vMange and Vsmart
- It ahs a public IP address and is deployed in DMZ
- It runs a virtual machine (can run on a router in smaller deployments)
- Multiple vBond orchestrators can be deployed with round robin DNS
- ZTP Zero Touch Provisioning service
- Cloud based shared service hosted by Cisco
- Utilized on first boot of vEdge router only
- Directs it to vBond to orchestrate joining it to the network.
- On Premises and Cloud
- vBond, vSmart and vMange can be deployed
- On premises
- Hosted in Cisco (or partner) cloud
- Most deployments are in the cloud
- vBond, vSmart and vMange can be deployed
- Building the Data Plane
- VSmart Controller directs the vEdge router to build a full mesh (default) of IPsec VPN tunnels between themsleves
- VSmart propagest policy and routing info to the vEdge routers with Overylay management Protocol (OMP)
- BF (Bidirection Forwading) VPN Tunnel Monitoring
- BF Detection packets are sent over all VPN tunnels
- this detects if a tunnel goes down, and also provides latency, jitter and loss statistics
- Traffic Forwading Option
- If multiple tunnels are available (MPLS and Internet) traffic can be load balanced over the tunnels
- Active/Active
- Weighted Active/Active
- Applicaation pinning Active/Standby
- Application Aware Routing
- If multiple tunnels are available (MPLS and Internet) traffic can be load balanced over the tunnels
- App Aware Routing
- BFD monitors the latency, jitter and loss across the VPNtunnels
- You can set minimum req. for an for an app with SLA classes
- SD-WAN ensures the app is sent over a link which meets its SLA
- By default traffic will fall back to another link if no suitable link is available.
- Traditional WAN Deployment
- Further Learning Resources
- Network programmability Basics video course: https://developer.cisco.com/video/net-prog-basics/
- DevNet Sandbox: https://devnetsandbox.cisco.com
- DevNet Learning Labs: https:developer.cisco.com/learning/
- Further Cert: Cisco Certified DevNet Associate
That was a rough section and there was a lot new stuff that I had to learn… Now time for practice test and lots of review.