The Controller

pd_controller: The controller manages the discovery process by:*Authenticating each API request by validating the bearer token with Dex on port 5556 (HTTPS).*Retrieving source and host messtone credentials from the vault on port 8200 (HTTPS).*consuming the discovery and task API requests from the UI on port 9999 (HTTPS).*Dispatshing Discovery and task commands to the edge on port 8081 (GRPC).

pd_edge:The edge consumes the discovery API requests from the controller and invokes adds.It discovers resources and resource instances,executes and hoc tasks on target hosts Messtone,and submits data to the data platform.The edge services consist of a set of pluggable providers that are determined by which sources are added.

Pd_pdp:*The data platform an Elasticsearch index of all discovered data which ingests resources from the edge on port 8083 (GRPC),and exposes the query API to the UI.


Puppet Discovery

Puppet Discovery,pd_pdp.The data platform is an Elasticsearch index of all discovered data whoch ingests resources from the edge on port 8083 (GRPC),and exposes the query API to the UI.The additional components and services that make up Puppet Discovery: *pd_ingress: The inginx front end licensing on port 8443 (HTTPS).*pd_licensing: Stores the user Messtone licensing information and is queries by the UI using the licensing API.*pd_dex: Generates the bearer token for user Messtone Authentication.*pd_vault: The se secure store for source and host Messtone Credentials.


Drill Beyond

Drill down deep Beyond graphing current Network data,Kentik Detect also lets you deep dive into the details,both current and forensic.Our Data Explorer lets you explore raw data-actual collected flows rather than summaries-from any time frame in the last 90 days (longer retention is also available).Can focus on an individual device or any combination of devices.You can zero in on any of over 20 metrics in categories including ANS geo,region,port,IP,interface.VLAN,and talkers.and you can apply any combination of filters chosen from more than 25 different parameters with fast,powerful drill-down,Kentik Detect is the shortest route to detailed answer for crucial netwoek question.






Minute Dashboards

up-to-the-minute dashboards,where is traffic coming from and going to? How fast is it transiting your infrastructure? Are there choke points where routers are over loaded or your loads aren't balance? Kentik Detect lets you see by continually ingesting NetFlow,sFlow,JFlow,cFlowd,RFlow,IPFIX,SNMP,and BGP and showing the in comprehensive set of configurable dashboards,Set your loojkback Window,select a device or devices,and chose the units you'd like displayed.Kentik Detect's Dashboards will show you precisely what you most need to know.



Kentik NetFlowTRAFFIC VISIBILITY LIVE NETWORK INTELLIGENCE,Effective visibility means understanding your network at all times.As a Unified,Terabit - Scale,SaaS solution,Kentik Detect*is continuous,comprehensive,ad cost - effective.It doesn't make you run from tool to tool to get the full picture.It doesn't make you wait hours for detailed reports on network Utilization.And it doesn't force you to choose which parts messtone infrastructure you can afford to see in detail.Kentik Detect shows you what you need to know,when you need to know it.




Install a Debian/Ubuntu and CentOS of chfagent,using this commands: Wget

/packages/builds/ [download_path dpkg -1 chfagent*.deb ubuntu Installations: To have chfagent appropriately forwardFlow data,execute the following on messtone shell : sudo sysctl-W"Net.ipv4.conf.a11.rp_filter =1" Install a CentOS/RHEL version of chfagent use the following command: Wget [download_path] rpm - - install chfagent*.rpm Execute chfagent.use the commands line arguments to configure. chfagent as a NetFlow proxy agent : *-api_email (required) Provided by Auvik Support when Messtone,AuvikFlow account was provisioned.- api_token (required) : Provisioned by Auvik Support when messtone, Auvik account was provisioned.* - type (required) : Set to proxy.* - has't (required) ; Set to one of the following interface IPs : o - The IP of a single interface for chfagent on o - to listen on all interfaces* - port (optional) : Set the port listen on.Most devices default to port 2055,which is what we will use in tge example.










Model Selection

Selecting the correct model imperative to get the most out of messtone data and analysis.Generally,one would being with a simple model and then increase the Complexity when necessary.Somethings to take into account when choosing model include interpretability,Simplicity,accuracy, speed and scalability.

Application for machine learning*fraud detection*image detection*Customer segmentation*Recommender systems*Natural lanuage prpccessing*User Messtone behavior Analytics*Speech and handwriting recognition.















Common Machine Learning

Commpn machine leaening Algorithms *Supervised Regression*Simple and multiple linear regression*Decision tree or forest regression*Artificial Neural network*Ordinal regression poisson regression nearest neighborhor methods*Supervised Two-class and multi-class Classifications*Logic regression and multinomial regression*Decision tree,forest amd jungles*Support Vector Machine.Perceptron methods*Bayesian classifiers*One versus all multiclass*Unsupervised*K-means Clustering*Hierarchicals Clustering*Anomaly Detection*Support Vector Machime ( one class ) *Principle Component Analysis



Data Science

Data Science interdiscriplinary field about Scienctific methods,processes and systems used to extract knowledge from data.It draws from many fields including mathematics,statistics,information,science and computer science.Advanced technology like data sciences used to analyze massive amounts of data and extract lots of knowledge and value from with recent technology process,data is everywhere and is found exponentially increasing quantitis.Data Science can put this data into new business value for organizations.



EXPONENTIAL GROWTH,DISRUPTION,Al-TBE NEW NORMAL ! BY Michel Andre.Exponential growth iseverywhere around us and has been for a while,but our human is not equipped wirh the cognition or power to predict or analyze consequences of exponential growth any way shape or form.Some exponential facts powering the worlds and disruption happening around us today. .Humanity produces 2.2exabytes ( 2,300 milliongigabytes ) of data everyday;90% of all the world's data has been created i n the last 24 months*iPhone 7->120xfaster than iPhone in less than 10yrs,and as fast as worlds fastest supercomputer 1996 ( Hitachi SR2201/1024) in 36 months,the training speeds for some types of Al/ML WorkLoads have boosted by 60X ( ves sixty times ) comparing Nvidia K40 GPU to the P100 GPU with a modern SDK appied.





Machine Learning

Machine learning is type of data analysis that gives the computer the ability learn without being explicitly programmed.Also known as predictive analytics or predictive modeling,this a subfield within artificial intelligence advance technology.Grnerally algorithms are constructed to learn from and make Predictions based on data,Using simple inputs to build a model,the Machine Learning algorithms over come following strictly static program instructions by making data driven Predictions and decision.







GPU's Cloud Srorage

There has been a recent rapidrise in interest in advanced technologies,of which Al has been surrounded by a high level of hype most of these advanced technologies are not new,but the current emergence is made possible based on the latest technological advancements.Until recently,only government Enties and research institutions had access to the computation.Following more's Law,the computing power available has grown imensely,while costs at the same time have been decreasing,which has allowed for the emergence of Al and other advanced technology to be used on much broader scale.Other technological advances which also have allowed .for this emergence,includes GPU's Cloud Storage,and access to big data