Duration 45:52
16+
Play
Video

Citrix Synergy TV - SYN225 - How to implement Intelligent Traffic Management for web app...

Steven Lyons
Product Manager at Mux
  • Video
  • Table of contents
  • Video
Citrix Synergy Atlanta 2019
May 22 2019, Atlanta, GA, United States
Citrix Synergy Atlanta 2019
Video
Citrix Synergy TV - SYN225 - How to implement Intelligent Traffic Management for web app...
Available
In cart
Free
Free
Free
Free
Free
Free
Add to favorites
331
I like 0
I dislike 0
Available
In cart
Free
Free
Free
Free
Free
Free
  • Description
  • Transcript
  • Discussion

About speakers

About the talk

Topic: IT

The internet is the blind spot in terms of delivering internet-facing applications. Citrix Intelligent Traffic Management delivers visibility and intelligent traffic steering for optimized user experience and performance. In this session, you will learn how implement Citrix ITM to gain visibility into your web app performance.Note: This session will be available for on-demand viewing post-event on Citrix Synergy TV.

Share

Okay, we're live Goods. Well, good afternoon everyone and thank you for joining us to hear a little bit about Citrix Intelligent Traffic management. 00:03 For me one of the most exciting technologies that we've introduced for a long time. Actually, I'm Jason pool on the from the product manager part of 00:14 marketing team and Steve Lyons from the product management team Tried Him and together. We represent the knowledge and the passion on itm. And today. 00:23 We hope to make you aware of some of the features functionalities and use cases that will hopefully enable us to impart some of the enthusiasm that we 00:33

have for the product on to you and you want to go home and start the straightaway 54 weekend. I'll just say that some it was nice to see that it was 00:42 included in the keynote in two instances, especially with the CVS in them delivering their big football game over itm that was made possible by 00:51 Citrix. So that was really good to see I'm a joke to tell me a little bit more about it as we go. Okay. So today we want to take you on a journey a 01:00 journey to delivering a better user experience to your users and customers it some I realize that some of this might be in you and so we're going to 01:09

do a little bit of a level set as we kick off and some of the problems with the challenges of our customers tell us they see every day and then we'll 01:19 spend a little bit of time discussing Intelligent Traffic management the product and the value so that you will be on a firm believer if you 01:28 understand what it is and how it works. Then you will find that use cases of the right for you rather than just relying on what we say is right for 01:37 you. We will cover some of that a little later on some of these cases that we see commonly with our customers today. But as I said if we make it 01:45

available in to your knowledge pool, then you'll be able to find new things to do with it. And I'm going to take you in a little bit of a voyage of 01:54 Discovery. We could present a thousand slides with demonstration. So Steve will be Guiding us through some of the some of the use cases on a customer 02:01 Journey from from a deployment that they have today to expanding to take over the globe and how citrus itm help them position their content in the 02:10 right place as well as helping them to deliver it in the right way and by then you should all be enthused and excited about it as we are and 02:19

then you're going to be say, how do we get hold of this we go to slide at the end of gives you three simple steps how you can go back to your business 02:29 today or tomorrow and set this up and get moving. Okay? so what are the things 02:36 that we're seeing more and more with our customers is that experience is a big Fran's differentiator. We all know this intuitively. We're bombarded by 02:46 it every day and in the it World customer experience in business success is frequently synonymous with application experience. And 02:55

that's what we're hearing a lot from our customers what we might want to think about even ourselves instant. Gratification is The New Normal 03:05 we can't blame the Millennials. It's been with me and my generation for a while and I'm sure it's going to carry on and get worse and worse as we go 03:15 the wheel use application on a daily basis. It's our job. We work with hundreds of them to her house everyday and we're waiting 03:21 for these applications to load to do their thing is not what we want to be doing you saw in the keynote how Maria had trouble and the the 03:31

context switching in the waiting was a problem for her. It was an issue and we want Try to get this situation away from you and Citrix and general and 03:41 also to you and how you deliver your business contacts contacts to your customers. And we seen also some statistics from 03:50 Gartner that we they think that in the next year or two customers today. They're willing to pay a little bit extra for a for a good customer 04:00 experience and they say that about 81% of businesses will be competing in the next couple years either wholly, or at least substantially on 04:09

customer experience. There was another one another I think it was a walker survey and they said the customer experience will overtake price and 04:19 product as the key brand differentiators. So this is something that is going to be very important to all of us as we go forward. 04:27 So how do we use it people improve that customer experience will traditionally reviews things like application delivery controllers to be 04:38 able to provide the right customer service provide the high availability of the application components improve that experience for all users 04:48

and that helps We've used application delivery controllers to optimize content for the return Journey to the client so 04:57 that the customer experiences improve there in that helps as well. We've also used as long as I've been in the industry Global server load balancing 05:07 by the users to the best available data center to get the concept that we want them to get and again that helps with three admirable things 05:15 application glittery controllers have been doing first and substantial. Of time, but as we move more and more of our workload into the cloud, so we 05:25

are more and more independent. On the internet as a delivery vehicle. And delivery across the internet is not always 05:35 straightforward. The internet is not a big white fluffy friendly Cloud. No matter how we choose to depict it rather. It's a seething mass 05:45 of different ISP networks competing with each other and peering as they transmit traffic and sometimes in a best effort to get that 05:55 traffic to your consumers and users and sometimes in so best effort to get that traffic to your users as you attempt to deliver that content. 06:05

This is not under the control of any single entity. Would that it were it would be easy to manage and user experience very very greatly. It 06:14 depends on the berries. Maybe the peering relationship that set up today might be different tomorrow. So you get huge Variety in latency 06:24 and availability and throughput and that effectively means the internet is a blind spot for application delivery. 06:34 And this is something that our customers tell us all the time. I know Steve's had a lot of PVCs with customers and they keep telling them we're 06:43

delivering stuff across the internet and we just don't know what happens. We get a customer calling us if they can't use our stuff. We look in our 06:50 data centers. And it's all green. So we just don't see that this blind spot is real and it affects all of our customers. Set 06:57 Enterprises. They desperately need something else. They want something better. If I think they deserve something better a solution with Intelligent 07:07 Traffic steering that has real-time deep visibility of what conditions are and the state on the internet. 07:17

And that my friends that is once I T M helps you do it helps you eliminate those it blind spots. It captures data about 07:27 the conditions on the internet and visualizes the state of the traffic between the user and your content and then uses 07:37 that same data to help optimize the delivery. I would get you use it to the right data centers overall so we can visualize 07:46 this condition and then we can use that same data to help optimize the delivery by sending you this to the best available destination. That is 07:56

the best destination for each individual users individual user at the time. They make the request. Because it might 08:06 not be consistent. That's the whole problem with the internet. It isn't consistent. And what is the best day to Center to go to get my stuff from 08:16 abc.com today may very well be a different thing from yesterday and another one tomorrow. But don't just take my word for it 08:24 IDC have recently published a report on this and they actually call this out as quite a pressing problem or issue a challenge 08:34

for the industry as a whole. They're telling us the application performance and user experience is going to depend dramatically on the internet State 08:43 and they tell us to find a way to cope with this an account for it and proactively adapt to it. Unless they 08:53 probably most interesting Lee. They tell us that the real payoff is not just looking at the data center and it's not just looking at the 09:03 state of the internet when you can combine both of those metrics. I come up with the best solution to send your uses to the most available the 09:13

best-performing and for want of a better word closest Internet site for your content to bring you use with the best experience. 09:22 And guess what? That's exactly what Citrix is done. Are Global server load balancing as a service? We delivers this today and it allows you to combine 09:33 the Deep knowledge of Internet State and combine that with the understanding that you already have within your data center 09:43 from your Citrix atc's or other third-party devices to feed information into Citrix it to enable more 09:53

informed and hence better decisions to send uses two more optimal sites. Who does give you this granular visibility of the internet the 10:02 analysis of the destinations and what that can do for your delivery as well as rapid and immediate actually respond to changes in either Gene 10:12 conditions in either the Datacenter or the internet conditions as well. So just to be clear this Global server load balancing is a service. This is a 10:22 combination of citrus itm Citrix ATC but it's shipping today. This is not a an announcement of a notification of intent to 10:31

supply this product gslb as a service. This is available from Citrix today. You can go home and get started with this now, okay. 10:41 All right, so I want to take a couple of minutes. Let's set the scene a little bit a couple minutes just to talk a little bit about the the the 10:51 products at and how it does what it does. Like I said to give you an understanding so that you can spot the use cases that will suit you. 10:59 And the first step to Intelligent Traffic steering is intelligent data and oh boy, the 11:08

Citrix collect a lot of data on the user experience on the internet. We collect data from about nine hundred million different user spread across 11:18 50,000 different ISP networks resulting in about 15 billion data points every day. In fact, you yourselves 11:28 are probably part of this data collection process. Whenever you visit one of our community member sites such as LinkedIn on Microsoft or Tumblr 11:38 you probably receive a small Java applet in your downloaded that webpage which one you finished browsing the page and it's quite on your computer. 11:47

Will Spring to life mask Citrix. Where do you want me to test today? And then what we will do is based on the or ISP Network in the data that we 11:55 collected so far we can sell you to go and check what's the journey time and the latency toy. What's the latency in the 3ft to Amazon in Sydney or 12:05 Azure in Ireland or softlayer in Dallas or something like that, meaning of themselves? These are small amounts of data 12:14 Clayton analyzing applied real learning to them. Then you can start to build this virtual map of Internet experience 12:23

which means that we're able to spot. Thousands of reach ability an outage issues every day. If 12:32 you check out our site i t m. Cloud.com live you will be able to go and see this is just a gift but you can see this live and see what the 12:42 conditions are on the internet across the globe and then drill into them and find out that CDN one is not available. For 12:52 users from Colombia and Brazil from these networks will give you that kind of granularity. I'm not kind of ground you allowed to eat then take that 13:02

and we apply that to all customers Datacenter traffic. So for example, if you're a customer and you have a website abc.com you may 13:10 have final 10 different points of presents and we will be able we will receive the request from your uses when they make their DNS requests for 13:20 abc.com and we will be able to tell them which is the best site to go to fool them at that particular time. I'm when they revisit tomorrow that 13:28 may have changed. We will know that in advance and then we will guide that traffic very intelligently. I'm out where this all happened. So I I 13:38

talk about briefly the story of how it happened. But actually Citrus itm is in fact composed of its own Some Cloud 13:48 hundreds of Pops of point of presence across various data centers across the entire Globe from Atlanta to Adelaide from 13:57 Chicago to Shanghai. And what does that mean? What that means is that unlike the old way of resolving traffic to your 14:07 data centers where maybe you have to travel all the way to the destination just to resolve the DNS address to find out that you need to go all the way 14:17

across the globe to make your request. Hundreds of Pops across the globe not only provide a more resilient and high-performance 14:26 infrastructure for your DNS and provide denial-of-service protection. It also means that your uses when they're coming to your site are never far 14:36 from a decision on a more rapid decision at that level means for a faster time the first fight and a better user experience. 14:46 Don't make sense make sense to me good is a lot of technical gobbledygook on this life for any cast network dedicated China Network hundreds of pops. 14:56

But really as Jason said what that means to you guys is it means business continuity it mean service availability and it means performance and so our 15:05 goal is to have the first Network that's really Enterprise ready to be able to do gslb is a service and we think we've delivered that I think so too. 15:15 You get excited. They genuinely get excited 15:24 with the technology because it is very exciting technology, but then they say, yeah, but but how's that for me? I've only got a single data center. 15:34

But there's value in this for all such as customers. No, thanks. Jason. Jason said 15:42 we really think there's value across the portfolio use cases for people using itm. We often see customers who come to us and say I have a single data 15:52 center. I have active passive data center. I don't need the sophisticated routing that you can provide. What's the value of this? And the truth of the 16:01 matter is is that when you only have a single data center, are you running I could pass it on to data centers or you have some simple Network you 16:11

actually need itm more than anything because you actually need to be able to optimize your performance in any way possible when your users are having 16:17 to Traverse back to your data centers to make DNS or gslb decisions are having to go a long way and physics that that's milliseconds that's time 16:26 that's performance that's impacted. So anyway, you can squeeze out performance improvements or reliability Improvement are especially important when 16:35 you have these simpler architectures where you don't have margin for error and so what can I can do for you can improve Performance and we'll dive 16:42

into some of these here in in a minute as we go but it can improve the performance improve the time to First byte that your users experience when they 16:50 do DNS in gslb resolve. It can can can secure infrastructure against DNS DDOS attack. We have we run DNS DDOS attack 16:57 infrastructure to help protect us and all of our customers and you get part of that protection is being on the I-10 Network reproduce management 17:07 overhead. You don't have to manage this service in the same way you do with boxes in your data center and you can prevent black holes that you often 17:16

see when you have less number of Pops to be able to handle some DNS issues that occur. So let's talk about these a little bit more individually. 17:24 So what you're seeing here is a benchmark of an Enterprise that is migrating to itm and we Benchmark their old infrastructure as well as the 17:34 itm at the structure that they're moving to the blue lines are the emphasis. Currently being run on Prime in the green lines are latency of itm. And 17:44 that orange line is the Improvement that you say sea by moving to ATM and you can see it quite significant So within the United States, you can get 17:53

improvements up 40 to 50 seconds. So about a 50% Improvement. But as you start moving out outside of the country where they this Enterprise actually 18:01 has already popped outside of the United States, but they still see significant improvements 300 milliseconds on the P95 you're seeing up to a 18:11 half a second and and 3/4 of a second or almost a second resolve times for customer. So that has direct impact on user experience and something like 18:21 itm can help you significantly improve the latency your customers experience and that's dollars and that's service availability. And so those are the 18:30

kind of performance improvements that you can expect from an infrastructure like ours. On the DDOS, you know for those who aren't exactly familiar how 18:39 that works. When you're running your infrastructure and users are accessing your services. You often times will see somebody attacking your 18:48 infrastructure and it's not necessarily somebody doing it on purpose. Sometimes it's inadvertent. Sometimes it's your own customers or your your own 18:55 developers and what happens when they do that as it will then take down your infrastructure. So if they attack your DNS infrastructure, you can see 19:02

the query levels of hundreds of thousands to millions of queries a second. So it's very hard to build that kind of infrastructure capacity on Prim. 19:09 And so that's why we have hundreds of pops and multiple anycast networks so that if you're running itm we can sit in the middle there and we can 19:18 protect you from these DDOS attacks because we're running store defense-in-depth that we have on our Network so built in D. Scrubbers blacklisting of 19:26 resolvers capacity baseball back all of these things that that sophisticated Services run that you would have to reproduce on your own if you were 19:35

trying to do this on Prim. On the management side know these are all the things some of the things that you need to do. If you're running the boxes on 19:43 Prem if you're running a DSL be serviced, you know, you're all probably very familiar with this you have to upgrade firmware rack and stack man is 19:52 connectivity. Aha pears pay people to do this monitor. The service is Monica the connectivity between the services you have primary DNS you need to 20:00 make sure that it's getting sent out to your secondaries. There's a long list of things before you can even configure your DNS and use it and with a 20:09

service like a man a service like itm. All you do is is configure you turn it on you can figure it that's it and then you can focus on something 20:16 that's more valuable for your business. So you can focus on differentiated will task. It's one of the greatest things that I hear from customers when 20:24 they tell us, you know, we don't get woken up in the middle of the night worrying about these maintenance issues anymore, or we can work on something 20:32 more important for a business rather than DNS and I think that's great. That means that we're making a difference for those customers. But let's say 20:38

you have more than one data center. You're not running active-passive you have multiple data centers, you run across clouds or multi-cloud you get 20:46 into sophisticated use cases. This is where I can really provides the greatest amount of value for you. This is where we can help you achieve the 20:53 business contact continuity that all customers are trying to achieve. So to make sure that your service is always available regardless of where 21:00 customers are accessing your service from as Jason was talking about the internet can be quite Wild and Woolly and it can be hard to understand the 21:07

the availability from inducer isps to your services. And where users are having trouble accessing and with the radar data where we map that and then 21:16 itm to used it to manage the traffic you can finally achieve that business continuity and do it in an automated way on an ongoing basis. And 21:26 also you can ensure the best possible user experience when you have more places you can send users you have more choice and more ways to trade off 21:36 those services. So whether it's deciding which data centers customer should go to your training off hybrid Cloud. So should I send them to my cloud or 21:43

should I send them to my data centers your part of the consideration you want to think about his performance and availability obviously, but they're 21:51 often kti's you might want to think about business focus kpi internal monitoring. You're doing a DC Health cost, you know 21:59 often times you want to make sure that when you're doing your your hybrid Cloud migration that you first sent as much traffic as it makes sense from a 22:07 performance perspective as much as possible to your on Prem data center so that you can optimize your Cloud spandex as openly as possible. You don't 22:17

want to have to just reproduce all of your on-prem and the cloud because user just started getting sprayed all over so you want to hear optimal about 22:26 how you distribute your traffic. And make sure that you're sending users to your data center when it was most optimal for you to do. So, so there's a 22:34 lot of cases where you can ensure the best possible user experience with itm intelligence and of course gaining visibility into the internet. We often 22:41 hear from customers will get calls. I love you to tweet from a user and they say and I looked at my monitoring everything looks great, you know, it 22:49

seems fine to me. But then I'm still hearing these issues radar data can allow you to see that. Maybe it's something else but you can also see are 22:57 there availability issues with the services are there is V issues and so at least you have a way that when your stakeholders call you and say why 23:05 can't I access your service you have another data source, you can look at and give them a good answer. So on the visibility size, it gets a dive into 23:12 a few of these items on the visibility side and inside of r r i t m portal. We provide you a lot of tools for slicing and dicing radar data and making 23:22

it useful for you. So you're able to geo-locate your. Your data centers in clouds and then see the performance of those from a latency an availability 23:31 perspective from across the world. So if your measuring your data centers with radar we can make that available to you based on your actual end-users. 23:39 If you're not we can give you her proxy measurements based on the rest of the measurements that we do so you can see how you know, approximately that 23:48 your users would be expected to perform. So it allows you to do a lot of what if scenarios and and Reporting which will get into a little bit in the 23:56

demo and then as well you can set up alerts when something happens. If you have a user population you especially care about you can set alerts on that 24:04 and be alerted to issues that occur and enter their number of ways you can use this for things like operational monitoring. Network planning or other 24:12 architecture planning where you want to see you know, where should I make improvements to my network based on how users are currently performing? 24:21 Jason Steve into that something which actually I think is one of the most exciting things about itm is the ability to ingest third-party data 24:29

when you want to make your decision, I talk about it as gsob the combination of a DC's and itm but it's not just restricted to Citrix ATC sure 24:38 we bring information from the Citrix ADC you spent the time and the effort and the money to put together great datacenters you want to be able to make 24:48 sure that you're utilizing the information in the knowledge that comes from those monitoring tools including the Citrus ATC. I'm bringing out into 24:58 your traffic management decision-making process. And that's what we do. We integrate with Citrix also other load balancers nginx and F5 25:07

note that I have to have two separate Integrations cuz it's not a single codebase, but it's not just a DC. So anything that you're using 25:16 for perhaps application performing Monitoring your data dog your cat point of whatever that is or whatever it is. You're using you can bring 25:26 information from that and bring it into the RTM tool so that you can then after that into your decision making so you may want to know you 25:35 may want to guide your traffic to Data Centers the where the application is performing better than where is performing worse simile. We integrate with 25:45

a lot of CDN and storage tools as well. So for example with CD ends, there is a cost associated with delivering things through content delivery 25:55 networks, and many of our customers are finding that it's worth the investment in Citrix itm to spread the load among many 26:04 different CDN providers as opposed to staying with just a single one. Restart it allows them to guide traffic based on. Well, it's only an extra 26:14 five milliseconds to go to see DNA which is a lot cheaper for me then CD and be so it's not a massive sacrificing performance, but 26:24

it's a good saving and cost you can start to do those kinds of calculation and start to work out what's going to be right for your 26:34 business from an economic perspective bringing this all together. So the beauty of citrus itm is it the 26:43 decisions can be made based on your business requirements. We looked at the sort of information that you can bring into it the the real use a 26:53 measurements of active and passive synthetic monitoring that built into Citrix itm in and out of itself as well as application health 27:02

or commercial aspects as well and the health of the platforms and service upon which all of this runs you've invested in. Can your Datacenter run very 27:12 well, you bought the best servers you bought the the fattest pipes out of the Datacenter. You don't really want the internet to be the thing that let 27:21 you down if you can help it. So that's where we bring all of this information together. You can bring out the business logic and start the design the 27:31 algorithm that you want to use for your content delivery. You're not restricted to the gsob algorithms that situation and other agency 27:39

vendors give you you designed the way that you want to do it you pick the CD and you want to use that's exactly what CBS did when they were 27:49 broadcasting the Super Bowl. They chose the CD ends that we're going to be best for the users but fit within their business model, and if something 27:59 was to break in the middle and make would automatically and seamlessly adjust users to different places. So you bring all that together and 28:08 spits out. Decision based on the metrics that you do that might be a DNS resolution with might be an HTTP 28:15

API call or something like that. But realistically it brings it all together and makes allows you to make decisions based on your business 28:25 requirements. We also built a platform to evolve with you as you make the journey into the cloud native application 28:33 architectures, and and you evolve your engineering practices are platform is fully API automatable so you can make it 28:42 so that as you do releases it if you know was updated with you and based on your your specific architecture and we have a lot of features around 28:52

helping you move to production in ways that are helpful to you and your team so you can get a lot of inside Saint why decisions are being made or 29:01 or for, you know are the availability reasons are for performance Surfer some third-party data reason so so that you have the transparency in the 29:11 understanding, you know, do I have a lot of users who are being being sent to a location because performances better or because actually there was an 29:19 outage in one of one of my keyring points and so my users got sent there by default. So we provide you a lot of insights or being able to Play a new 29:28

traffic management role to a small percentage of your populate population to validate that it's behaving as you expect before you roll it out to the 29:37 rest to all of your production traffic. So all of these things are built to help you integrated into your you know, the new one involving 29:45 architectures, but also with the fundamental thing to help you improve the business continuity that you have proved that applications are performing 29:52 the way you want. And if you're there traffic management stop them all for you. Okay, so I think that we we talked a lot and I can see and feel 30:01

the enthusiasm and excitement in the room and it's building. I think it's time now to switch over to some some demonstrations. It's just to remind you 30:11 what we've got here is a customer who's got an active passive setup. I think in the US and they're looking to expand we're going to show you how we 30:20 can use on performance visualizer to help him select the best sites to position their new content. And then how we're going to help them deliver that 30:30 content most effectively utilizing there a DC and itm together Steve and so he said we're going to look at how a customer may look at this 30:39

how you guys can look at this when you're thinking about hosting your web applications or your work spaces are other things. So the first thing to 30:49 look at is is this is a tool we have our performance visualizer. It's publicly available. You can go to itm. Cloud cam performance and we have data on 30:56 all of the major Cloud regions and and others Message me Benchmark and you can show you by choosing the ones you want what the performance the latency 31:04 performance is for these different services. So in this case, I have Azure Cloud you at least you can see you over here on the right. So I'm using in 31:13

this particular example USD east and u.s. West and currently running an active passive architecture. So might u.s. East is is currently 31:21 running my traffic. So this is what I can see the actual latency that my customers would experience for an HTC at DNS and HTTP request. You can see in 31:31 the Eastern Seaboard. It's not too bad. It's it's decent latency, but when I get anywhere outside of that it starts to decline rapidly and I have 31:41 about a 307 millisecond Global latency, that would be expected. So my users are as I expand worldwide some companies like say 31:48

Citrix Enterprise or many of yours, we have customers and users and in Europe and Asia and India and other places and so will I You to look at that if 31:58 you're thinking about increasing the size of your architecture, it's a Well, I know I have some users in Ireland. So I want to to add them and and see 32:08 how that impacts my latency profile and then in his meows and I didn't say Tokyo or over in Asia and see what the 32:16 profile is there. One of the things you notice as you're looking at. This is that performance didn't actually always increase as I added new locations 32:26

to my architecture. And the reason is is that currently I'm I'm looking at a round-robin optimization that some customers use and what happens is than 32:35 users gets sprayed all over the world when you're when they come in and so it actually can decrease the user experience for some users. So that's why 32:45 you want to do something like Geo proximity and then you can see it's a little bit better shows me, you know where I want to post my users and so you 32:55 can see some improvement there that I might want to look at and then also I have users in in India. So I might want to think about what the 33:04

performance of those users are going to be and you start seeing some interesting things with data as you start playing with his data. So for example 33:11 in India, if I have users in India, I often need a pop inside of India because of the way that India is. Outside of its neighboring countries. 33:18 It's hard for users to get a good experience unless you're actually in country because of the peering and other things that can significantly impacted 33:28 performance. So this is the kind of thing as you look through the data you can find and learning is for you as you go and so so 33:35

I'll leave that off for a moment just to appoint as your guys you guys are using this tool and then you know, you also need to think about not only go 33:43 proximity but using something like itm where you can optimize your experience, even further improve the performance of your your DNS resolution and 33:53 make smarter choices, then you can improve your performance even further and lower your average response time now, Down to under 200 from where you 34:01 started before. So the Geo proximity is kind of what you would have with regular gslb classic vs will be that we collected from 34:09

every eyes pee in 900 million users and so on and so forth. That's right. Now we're going to show you the next steps. So I've decided for my customers 34:19 that I want to place data centers in Tokyo and Ireland. Is it in addition to my existing East and West coasts Datacenter some West Coast. I'm going to 34:28 continue to do active passive. I've decided I want to leave that as a and active passive within the United States for business reasons. So I've gone 34:36 to those we've gone to those pots that I have we have in that case. I'm running as you're here in the US and then AWS outside of it and we measure 34:45

both of those and make them available. So I've created those pops and now I'm going to let you know when I've added those netscaler to those pots that 34:54 generated the settings that I have their ready to go live. So now what I'm going to do is I'm going to import the configuration from my net scalars 35:03 that I have worldwide and import those into itm so that I can start setting them up for traffic management. So This Is A & N S 35:10 cam file that just has my my Global load balancers in it. And so then what item does is import this in I created my two new. 35:20

Cslb locations and now I can do similar kind of things to what we did outside. I can start playing around and answer to looking at how that impacts my 35:31 performance start getting on going reports on the performance of of my data centers. How am I impacted by the different routing rules that I use very 35:39 specifically in two countries. I can dig down what is my latency profile inside of those countries? And so now I'm ready to start building my traffic 35:48 management Logic for for these services are more than just a country. Can you dive deeper into different networks will providers and stuff with this? 35:57

Yeah. Absolutely. So we look at data is p e r a s n level so an ISP and and Country level they're down to so you can look at it as sort of a 36:05 continent country region state so you can get a place as local as networking States. You can say, Oregon. I'm in Oregon on Comcast. What is 36:15 my latency profile for my user's latency profile to my different data centers. I am in. Time Warner in New York yet. Not sure who Atlanta use it so I 36:25 could be like that. So I imagined Coca-Cola probably has all the way. I have access to those. 36:35

So now I'm going to show you how you can start setting up the traffic management. Which again is it is very simple. So here's another account that I 36:44 have the symbol of the same data gslb that I had before selected. And what I'm going to do is I'm going to create a few traffic management application 36:52 so you can show you guys how simple this is. So the first thing that I'm going to do is I'm very quickly going to create a optimal rttr 37:00 performance-based optimization doing this case. I'm going to send users to wherever it is. That is the fastest for them as measured by a 37:10

radar. So I'm just going to run to this right here because it's so fast that I can do it. right now so I 37:20 will do this. So what I do is I chose optimal rtt. For the which means the fastest performance and then I'm going to choose 37:30 all of the the difference. Yes, I'll be is that I have just created and give them their they're see 37:40 names that I've that I've created for them. And then I added each of those three that I have. So my AP EU and US 37:49 data centers here. I can set waiting if I even though I want to choose performance. Maybe I want to wait it a little bit more to the US ER or 37:59

somewhere or I can assign specific Rules by country. So perhaps in Germany, I want to go to the EU pop because of you know, David data sovereignty 38:08 reasons are or whatever you can easily do that here in the configuration and I clicked done and now it's been created as is easy 38:17 as that. I now have a performance-based optimization for my customers. I can use this in my DNS and very simply can I get a performance-based 38:27 optimization? So it's as simple as that that's all you have to do to deploy this application into your infrastructure. Now let's say that I have done 38:35

that and now I want to get a little bit further down into the sophistication. So I want to ingest some metrics from my ABC's to make decisions on how 38:44 I want to route my traffic so it could be that the fastest available data center is not actually where I want to send users because because you know, 38:53 baby I have some of my web servers aren't actually available behind my my ABCs. So we're going to do something a little bit more sophisticated in this 39:02 case and will show you how how simple it is. So here I've already created a nap and here's how it works. I'll show you what it looks like. I created 39:10

one of my traffic management rules and let me pull it up here. And in this case, we allow you to specify your criteria that you're using with 39:19 a Javascript app that runs on the server. So it provides you really good flexibility for how you want to Define your criteria for traffic management. 39:29 Can you can see the same data centers that I used before and what I'm going to look at is I want to 39:36 to look at the health data that comes from my from my ABCs. So we allow you to bring in this third-party data as Jason and 39:46

talked about you can bring in through what we call Fusion you can bring in data that you choose to make available to the traffic manager. So in this 39:56 case, we have all of this different data available. I'm going to choose one that I want to use. I want to use the vs lb Health, but again any of this 40:04 is available, so we hear from customers and some cases they wanted you at use SSL termination errors or health scores or just random data that day is 40:12 important to their organization that we haven't currently built implementations for they can do that via something like this. So inside of this custom 40:21

map I use this vs lb health And here is a shows you what one of our responses look like. So for Atlanta where I'm I'm 40:30 calling my application or calling my traffic routing rules you get as you'd expect the US response for where to send me so I'm at in Atlanta. And 40:40 so it says the US for you is the closest so obviously that's where you want to be sent. You guys are probably like well that's kind of obvious. Okay. 40:49 So in the case here is one of my data centers. I'm running to web servers. I'm going to disable one of those web servers. And first I'm going to login 40:57

to the demo God's always do this. The reason that it showed up that the US actually looking at in this is always pick the 41:06 closest unless the server Health behind it as we saw was not a hundred percent. So what that means is we're going to fail one of the servers behind 41:16 the netscaler behind the Citrix ADC. I'm sorry and then that will reduce it to 50% We got two behind each one and then in theory it. It should say 41:26 that's not the site within a remove that from the pool. We're going to send you to the next closest site. All right, so I'm going to disable that. 41:33

I'm going to give it a little bit for give it a minute or two to four that to propagate and so in the meantime, we're getting towards the end. So 41:43 Jameis Jason, do you want to talk a little bit about the wrap up while we wait while we wait for this? So what I want to say to you is that this is 41:51 real. This is you're going to see the cement does taking it. We're doing a whole rather than the other way the power of the 41:58 combined ATC and itm. I think Madeleine Albright said this about Bill Clinton earlier, it's snowing the the domestic and 42:08

foreign policies that you want to put together and how they interact and that is exactly what Citrix itm and 80's you doing here together to create 42:18 this gslb as a service. Couple slides outfit is that now work? So we're now it's still going to the u.s. Give it a second. I 42:27 say we just updating this every minute and I remember that long thing that we had all the different date of these adjust Nitro feeds that we're using 42:37 so you're familiar with ATC then you should be familiar with all of the information that you can get the thousands of different things that you can 42:45

get out of it and use any of that in your decision-making policy the EU server, so I'm now 42:51 going to the second best server for myself and it automatically did that. We didn't have to take it out of rotation and importantly it will tell you 43:01 why that is so here inside of your your application at now tells me that you know, my vs vs lb health is 50. It's not a hundred. So it 43:09 tells me not only that it take took it out. It tells me why it took it out so that I can go deal with that problem. So you can see the power of the 43:19

integration of gslb and itm. Back to this is the power of the integration of gso be an IT. And remember it's about a holistic approach to 43:26 your application delivery getting to use it to the right place that's step one using the right atc's within the data center to optimize the traffic 43:36 routing in the Datacenter and the content return to the user that's steps 2 and 3 as well as securing it for stuff for now. Don't take our word for 43:45 it. This is not a new product. We got lots of customers using this you heard two of them yesterday. We've got a lot of others I did mention Microsoft 43:55

and Linkedin but it's not just you don't find those under rocks everyday, but you know, there's a lot of other customers in different areas and 44:03 different verticals that are using this as we speak. Okay, so Steve, how do they get started with this three simple steps, right? Yep. All you have to 44:12 do log into to citrus. Cloud cam. If you have an account already go ahead and log in with that. If you don't have one create one, all you have to do 44:21 is go there's a tile in citrus cloud called Intelligent Traffic manager. Just click on that and then it will add it to your services. You'll drop you 44:30

right into itm and then you can just start using it. So it's free to use food and access the data and then there's a a trial you can access the the 44:38 traffic management portion so you can get started right away on your own. Come and see us after it's okay just before you leave a couple of things. I 44:47 can feel that excitement in the room and I want you to amplify that in social media. So feel free to tweet about the session. This has been to 25. 44:57 I've been Jason Poole. This is Steven Lyons. If you've liked it tell the world if you didn't like it come and tell us and how we can make it better. I 45:04

just a couple of other things before you leave. It'll be another session. Mitm tomorrow afternoon session 718 we will be talking about things a little 45:12 more specifically around the delivery of video with finding a lot of customers actually deliver a lot of the video. I'm through CD and so on and so 45:22 forth. Okay, and one last thing, please rate this session and continue with the game on and let us know what you think. Thank you very much. Thanks. 45:30

Cackle comments for the website

Buy this talk

Access to the talk “Citrix Synergy TV - SYN225 - How to implement Intelligent Traffic Management for web app...”
Available
In cart
Free
Free
Free
Free
Free
Free

Video

Get access to all videos “Citrix Synergy Atlanta 2019”
Available
In cart
Free
Free
Free
Free
Free
Free
Ticket

Interested in topic “IT”?

You might be interested in videos from this event

September 28 2018
Moscow
16
122
app store, apps, development, google play, mobile, soft

Buy this video

Video

Access to the talk “Citrix Synergy TV - SYN225 - How to implement Intelligent Traffic Management for web app...”
Available
In cart
Free
Free
Free
Free
Free
Free

Conference Cast

With ConferenceCast.tv, you get access to our library of the world's best conference talks.

Conference Cast
505 conferences
19653 speakers
7164 hours of content