Recently I got an opportunity to work with a product called “ExtraHop“. If you are already familiar with that product then you may not want read this post to realize its value. If you haven’t come across this product yet, then please read further and get an idea how can it help you in your network environment.
We got this tool as a replacement of our Netflow monitoring tool. But it has much more capability than simple netflow monitoring. Primarily it is a big data analytics platform that can help you with following. It is a very powerful tool for
- Network/Application Performance monitoring
- Network Security ( Threat Analysis, Forensic investigation/Machine learning)
- IT Operations (Real time analytics, Network Baseline)
ExtraHop solution can include different components (it can be physical appliances or virtual servers)
- EDA – ExtraHop Discover Appliance (Top level application monitoring-metadata)
- EXA – ExtraHop eXplore Appliance (for transaction level details)
- ETA – ExtraHop Trace Appliance (for packet captures)
- ECA – ExtraHop Command Appliance (management appliance)
Below diagram shows how these components interact with each otherYou do not require all of those components to start with. EDA is the primary component that you should have to begin with. It can be deploy as a standalone unit or you can integrate it with a command appliance (ECA), usually a VM. With ECA, you can have multiple EDAs, but still you can manage all via ECA. Here are different platform options available.
Mainly EDA get data feed from two different sources
- Wire data ( from SPAN/RSPAN/ERSPAN)
- Flow data ( Netflow data from network devices/servers/etc)
Below diagram summarize the capability of this wire data with ExtraHop. It does not store all the raw-data, so you will keep required meta-data and provide deep analytics on wire speed.
Here is the protocol breakdown that it can analyze. Those are highlighted are paid modules if you want to get visibility of those protocols.
Once you go to Metric > Sources > Applications > All Activity, you can get a summarized view of all your wiredata that you sending to EDA. That a great way to get an overview of each of those applications.
Having this sort of visibility of your network will give you enormous value. You do not want to install any agents on application servers, any polling mechanisms or traps from those systems (like SNMP, Syslogs) to get that visibility. Here is a summary view for NAS performance that you can choose from “Activity” section breakdown.
System will automatically group based on activity (Metrics > Groups > Activity Groups) , so you can easily find out those applications you want to have a performance monitored. You can easily navigate to a system that you want to know about from there as well.
If you have specific server performance to monitor, then you can filter that by navigating to Metric > Sources > Devices & then search for specific server. I used one of CMX server to monitor its activity.
As you can see from overview page, you can easily find application throughput, what are the top protocol used by this applications. It is a great way to know your environment how these applications work. Also it will classify “server” & “client” activity of those protocol monitored by ExtraHop
If you click activity map, you will get a snapshot of what other systems is communicating with given CMX. In my example CMX is communicating with WLCs suing SNMP, SSL where as it use TCP & SSL to communicate with Splunk server. Size of those circles indicate traffic volumes (so you can easily identify what to focus based on traffic volumes)
You can create your own dashboards & share it with user who manage those applications. To get an overall view of your environment, you can go to system level dashboards. It has two views, “Network” & “Activity” dashboards.
Here is what you see in Network dashboard
From the Activity dashboard, you can get an idea of top applications and summary of each application protocol. This is really useful to know your baseline and help you to quickly identify any anomalies of your traffic patterns.
This product has some other advantages too. Since you can access these metadata easily, you can develop your own bundles to monitor custom applications. Here is a link to already available bundles.
If you go to Metrics > Networks section, you can see all devices that sending netflow data on to it. Then you can select required interface traffic to be monitored.
Below shows traffic distribution of WLCs.
With respect to wireless netflow, currently, ExtraHop is not supporting all the fields of Netflow traffic send by Cisco WLCs (As detailed in this post). I would like to work with them and get it improved. If they do, below details can be username instead of IP address (in wireless given user device can assign different IPs, so stats per IP does not make much sense)
If you have EXA appliance (VM or Physical) you can get transaction level details. “Records” option highlighted gives that detail. Below example shows it can see specific SELECT query run on a database server by a client
It also got some advanced capability such as anomaly detection based on machine learning (This feature is called Addy, Refer this packet pusher podcast for more details how it works). Basically it will detect anomolies and alert you, in the background it will take some intelligent decisions & respond differently if same alert keep continuing (eg it may disregard as network admins not treating it genuine).
You can use “Alert” view on GUI to see those anomalies. Here is an example
As you can see, it is a very powerful tool that will help you to get real visibility of your network and assist you keep it secure. You can register for a demo by using this link.
My post got Extrahop’s attention and they have published a post on their blog about our experience. Please read it from here: