Setting the Stage for End-to-End Visibility with Riverbed

well hello everybody my name is phil gervasi senior technical evangelist with riverbed now what i'm going to do for you today is set the stage for riverbed's end-to-end visibility story so when we were talking about networking field day we thought to ourselves what would be the best way to present to the networking field a26 delegates the full story of how riverbed's end-to-end visibility solution actually works how it operates full fidelity with the depth and breadth from the client to the cloud well one option is to put together 100 maybe 150 powerpoint slides with boxes inside of boxes and lines connecting them that would work right you could get some the gist of our features and we'll go through that like a typical presentation but we thought to ourselves nah why don't we just show you how it works in a real scenario that we developed for you in a demo environment so it's all real data and uh and at later in the presentation i'm going to show you if you want to roll your sleeves up and get your hands dirty how you can access the same demo and go through the same steps that we're going to go with through today so what we're going to do presentation by presentation is walk through a single troubleshooting scenario just like a knock would in an enterprise organization with a help desk a level one a security knock and and maybe a level three engineer now a little caveat all the demos uh rather the screen captures they're all recorded so you can certainly stop the presenter and ask a question as you go we welcome that we encourage that but they're not going to be able to click around to other things uh we did focus specifically on the scenario which means that we are going to skip over certain features because we uh you know to go over app response for example would take four days we're going to say focus on this troubleshooting scenario to show in real time how you can use riverbeds end-to-end visibility solution to solve an application performance problem so let's get into it and my slides are not advancing there we go steel demo rental cars is a national us-based rental car company uh located across the us in the west region central region and east region this is our company that we're going to be using as our demo today now if you're wondering where steel demo came from that's a play on words from riverbeds steel uh steelhead appliance seal central all those kind of products so we have san jose and la in the west region dallas fort worth in chicago in the central region and in the east region jfk which is near my hometown and atlanta now if you can picture in your mind's eye being at one of these airports or a similar airport there's a rental car kiosk with a customer service agent operating in this case a windows computer helping folks to check in reserve a vehicle make a payment get their keys and get them into a vehicle as quickly as possible now here is the line of business application that we're using rentalapp.steeldemo.cloud and you can see here this is a web application where you can you know choose from a drop down to pick up your location a pickup location pick up rather dropoff location our employees log in via this same web application but we also use this with our external customers and if i scroll down here i'll show you what i mean we promote uh various promotions throughout the year like fifty percent off for new members um so it's very important and we can also you can browse apparently mercedes-benz r3s although i don't think those are all mercedes-benz r3s and they're only 99 a day so that's a really good deal but this is our line of business application rental app.steeldemo.cloud used by our internal folks and external customers alike help desk tickets now are coming in in fact the first one is from jfk we went with jfk because again that's my hometown but we're getting tickets from across the country that rental app is running slow on those windows computers at the airport service desks now what's happening it's taking a long time for rental app the website to respond it freezes the logins for employees to log in with their company credentials very slow and sometimes payment submissions fail altogether now understand this the customer lines are getting longer it's taking a long time to get somebody from getting in a line to getting in a car and here's the thing steal demo rental cars we're going to rent you a toyota corolla for 99 a day maybe that's too expensive but whatever our our neighbor at the airport is our competitor and they're gonna they're gonna rent you the same toyota corolla for the same price pretty much minus a dollar or two so our business model is customer service getting you through that line as quickly as possible get the keys in your hand and get you out the door make it as painless as possible but it gets worse it's not just internal users we have a customer complaints line we wish we didn't have one but we have one and customers outside our uh internal users are calling up saying hey i'm at my house on my windows computer i'm on my smartphone using your app and it's running slow it's not working properly and some customers have reported even giving up in frustration and going with a different rental company this is money we are now bleeding we need to solve this problem as soon as possible so let's get into the topology this on your screen here is the steel demo rental cars network it's a multi-tier application that you can see with an sql backend in our private data center we have a three region web front end and we use uh three regions simply because we want to load balance we just use dns load balancing we load balance uh uh where people which particular uh web server they go to so if you're in la at that airport you go to the west region if you're in jfk you go to the east region we use that same load balancing for our external customers too all in the effort to give them the best performance the best experience possible that's the key here so if you see here in the lower left we have our airports and a real company might have 600 airports not or you know 600 locations not six but you get the idea and they are um using the company sd-wan and you can see our internal dns that we host in azure our external customers that's way on the right i have a couple smartphones there they're using the public internet and whatever external dns they use what we've also done is set up and this is something that i've done as a network engineer many times in my in my career uh we just threw a cheap broadband modem in a west coast airport let's say la central airport chicago and jfk in the server room a cheap broadband modem and a test computer behind it it's not running anything there's no company stuff on it but we use it to test the website the performance of the website to simulate what external customers performance is like we want to know that as the i.t department so you can see here here's our topology here's our layout the little a on the little computer screens for our airport computers symbolizes the eternity agent something that john hodgson is going to talk about in a little bit as when he talks about visibility from the client perspective the rs that are all over those are various not all the same but various visibility tools distributed throughout the network all together giving you that end-to-end client to cloud visibility so um i'm going to start this off i am going to take the role of our level 1 help desk but i'm not going to roll up my sleeves and start looking at packets and that's not my that's not my role so what i'm going to do is start with portal now as we go through these presentations you might start to get the feeling hey you know some of these things feel like you know i'm looking at a series of point solutions that's not at all the case i want you to look at all these solutions as one or the individual tools as one end-to-end visibility solution that all roll up into this single dashboard portal so for in portal we have say you know desktop metrics uh container metrics server metrics vm metrics all coming from eternity feeding into portal we have our synthetic tests and you know traditional device information that we gather via snmp from netim feeding into portal we have tcp analysis tcp connections analysis from app response feeding into portal and so on application flows with netflow it all feeds into portal and then is presented to you in customized dashboards that you customize based on who you are so if you're a developer a network admin a security admin whatever maybe you want maybe you're the voice over ip administrator for your entire company you just want to see voice gateways you know packet drops that's what you customize portal for so enough with the slides let me show you what portal looks like and let's start troubleshooting the scenario this is the rental app performance health dashboard which i created and this is uh this is real data by the way we have real web servers up in aws east central and west uh and they are really performing poorly we really have those airport computers all spit up so this is real data and already i see as a level one my goodness we have a problem in rental app east region that's a that represents an http status test now i configured the threshold me personally i configured that threshold to say you know when a certain number of http failures occur throw up a critical alert you know you can adjust that threshold the same thing in the central region we have a warning i don't know how many that is i don't remember what the threshold was set at but we have a warning there have been some failures it looks like the west region region is humming along notice we have response times which um i think vince will talk about what response times actually are a little later i'm level one so i want to see a little of everything but this could all be one geographic area it could be all your web application it could be just network devices here i have the airport sd-wan devices you can see i have atlanta uh dallas-fort worth jfk those are the top three everything as far as cpu utilization it's running fine it's in the green uh that i could also add memory utilization link utilization whatever you want to see on the right i have aws uh net network metrics and those that would be thing like re-transmission delays you want to know that kind of stuff what's going on in the network uh between uh between wherever my client is and aws let me scroll down a little bit and here you're going to see a few more graphs you can have as as complex or as simple as you want by the way our west region is in milliseconds and it looks like our response times are two milliseconds little under that's fine that's great we have nothing in the west uh the central region i don't know what that's all about i mean we're gonna have to troubleshoot that that i expect to see something there in our east region notice the that's in seconds it's so bad that we skipped milliseconds and we're right to seconds we have a spike over 15 full seconds looks like the average is between five and ten that's that's really bad we have a very bad response time our top applications you can see rental app being the the majority of the traffic which i expect you can see i have the external machine test metrics there and server turns that's how many times it takes to go back and forth but ultimately this is what portal is all of this information that you see here all of that is being fed by the variety of tools that we're going to now look into over the next few presentations and keep that in mind you might actually be able to solve your problem from portal alone right you could see okay we have something going on and uh it's pretty clear what the issue is but what happens when you need to go deeper and say this is a complex problem we have a complex network multi-tier application overlays sd-wan all these things going on i want to check my database connections on the backend as well well you need to have like vince said that depth and breadth and each one of our tools is focused on a particular sliver because it's a specialized device ultimately always feeding into into portal where we started today so i don't know what the problem is i see a couple clues what i'm going to do now and conclude by saying we use portal to start off because it collects all that information from all our tools to get a quick high level view of what's going on going on with rental app we kind of got a feel that the east region is not so great and the network might be okay but that's really all i am or rather all uh all i know is a level one so what i'm going to do now though is hand it off to john hodgson who's going to look at this remember this uh end-to-end visibility from the client to the cloud he's going to start as kind of an endpoint engineer from the client perspective thanks very much

Channel: Tech Field Day

Views: 101

Rating: 5 out of 5

Keywords: Tech Field Day, Gestalt IT

Id: l-ZFyq-sJsQ

Length: 11min 50sec (710 seconds)

Published: Fri Sep 17 2021