Duration 41:32
16+
Play
Video

Citrix Synergy TV - SYN146 - #RealWorldExample: SD-WAN eliminates the consequences of network...

Valerie DiMartino
Sr. Product Marketing Manager, SD-WAN at Citrix
  • Video
  • Table of contents
  • Video
Citrix Synergy Atlanta 2019
May 22 2019, Atlanta, GA, United States
Citrix Synergy Atlanta 2019
Video
Citrix Synergy TV - SYN146 - #RealWorldExample: SD-WAN eliminates the consequences of network...
Available
In cart
Free
Free
Free
Free
Free
Free
Add to favorites
482
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

SD-WAN helps all types of geographically distributed businesses create reliable networks that eliminate downtime and lost revenue. In this session, you will learn firsthand from a Citrix customer how Citrix SD-WAN can ensure that the business will not be impacted by network outages, insufficient bandwidth or lack of network visibility. Make dropped VoIP calls, loss of Citrix sessions and poor video quality a thing of the past. Learn how Northside Hospital in Atlanta centralized the management of their network environment, consolidated their WAN edge infrastructure, added redundancy and doubled their bandwidth, saving thousands of dollars. If your company has a distributed workforce, come to this session to learn how the unique SD-WAN technology ensures the best user experience for video and virtual desktop sessions, real-time communications, and more.Note: This session will be available for on-demand viewing post-event on Citrix Synergy TV.

Share

Good morning, everyone. Welcome. Thank you so much for coming. This is sin 1 4-6 real world example, we're going to talk about sun with Northside 00:04 Hospital. So so glad to have you here. I'm in SC product marketing and joining me. Today is Dave words. He's network network architect at Northside 00:12 Hospital. So we're going to get into a lot of his environment as well. So and I'm going to give you some scoop on what s u n is doing these days. 00:22 And I just legalized to let you know if anything's forward-looking is subject to change as far as when it comes out. And then 00:34

I'm going to have you and just quickly talk about Northside get us started. 00:44 Quite a few. All right. So Northside Hospital, we're obviously we service the Northside of Atlanta Destiny Northside. We do have some sites for South 00:53 in Atlanta, but we are known for one thing. They call us the baby Factory. So On Any Given year, we average about 27,000 newborn children born at 01:02 Northside. So it's about 75 per day across our three Hospital campuses wore. The number one Hospital work for in Atlanta or in the top 50 in the US. I 01:10

think the number 52 to work for in the United States overall not just hospital but employers got about a hundred 50 sites that is broken up into four 01:20 sides acute care and large Medical Office Buildings. So if you care just the main Hospital's the bulk of our sites are these ambulatory side so 01:29 Radiology clinician site Farm pharmacies remote pharmacies physician practices, and we're in the process of acquiring if you're from Atlanta Gwinnett 01:37 Medical which has another two hospitals within their system with two acute care systems and another 30-plus remote site. I'm going to 15004 East 01:46

thousand beds and we see about 3 million people a year currently within our three hospital system. I don't know what the numbers going to jump to when 01:55 we finally merge with Gwinnett that's been an ongoing acquisition. Great. And so today we're going to talk about the healthcare landscape. It 02:02 also applies to other verticals as well. Dave is going to talk about what's going on in his Network. I'm going to talk to you about some of the 02:12 changes to the network in general and what we're seeing with our customers today the transformation that they're undergoing as far as digital goes 02:20

what Sun can do for these color is doing for these customers and What Makes Us sort of unique so Without further Ado 02:29 in healthcare and another verticals. There's many many factors that are affecting the network and it from 02:39 offering new services to being always available to moving to the cloud with in healthcare. There's now accountable care organizations. There's 02:49 value-based care. There's iot communities of care people are able to now on use devices to check their Healthcare at home 02:58

and and infrastructures being Consolidated as well. So there's a lot of complexity that's being created because of all these changes 03:07 and I've actually highlighted summon green here for Dave to to sort a comment on the m&a and onboarding 03:16 find Rose at a ridiculous, right? I think we acquire anywhere between 2 and 3 new practices per month. So is a network engineer in a network engineer 03:25 team. This becomes a huge disrupter for us. You know, we were cruising along doing our thing. And then we got 30 days that site has to go live and oh 03:35

by the way 2 days later. This other sites got to go to the bring us to win in for us has just saved us so many times because we can now meet these 03:42 timeline. You can't get a circuit in 30 days from any major carrier circus we can get him in seven days. This light goes live. We don't have to worry 03:52 about it. And then obviously with some of things were going to discuss here with us. You win single-circuit SD when I still have a good quality of 04:00 experience for my user until I can bring redundant circuits and then back up, you know, LTE things like that. Okay. So in terms of healthcare 04:07

challenges, he's obviously my span across different verticals as well. So Network complexity lack of flexibility obviously 04:16 mpls has been around for a while. So that's something that's not very flexible. Whereas, you know, broadband and things like that are much more 04:26 flexible than at work is costly to manage and and upgrade. There's things that are driving more 04:35 demand more high-performance applications voice vdi 3D Graphics Imaging modeling that are high fees 04:45

really require better links more bandwidth and and More prioritization in general brownouts. So linked conditions without 04:54 something detect knows they might not be you might not be aware that those kinds of things are happening with congestion or Jitter but you know, your 05:04 users might be having a poor experience, but you know, there is peace today. You might only know if it actually your link actually has a full blackout 05:14 so And then you know application for 4 minutes. Obviously with Healthcare EMR has have to be always available the move too 05:22

fast applications that's going to happen at some point for everybody virtual apps and desktops. It's the best way to really 05:32 consolidate the delivery of the applications to the users in different locations and obviously clinical tools that swell and then stops is in a 05:42 variety of different locations for rehab facilities to clinics to those large mov 05:52 learned about today, you know doctors are moving around Health Care clinicians are moving around and they need access to the patient records at all 06:00

times. They've got to be able to to get the records and the Imaging back and forth from the data centers back to to treat the patients. 06:10 And so when we look at the wheel and you know, obviously it's really about the experience today. It's not so much about 06:20 the you know, that the network links and things like that in the speeds and feeds it's really about the user experience that we want to improve mpls 06:28 has been really dominating for the last twenty or so years. We don't see it going away necessarily, but we do see that we don't need to just 06:37

depend on that slowly and put all our eggs in that one basket. We can actually complimented with cheaper more available bandwidth and still have that 06:47 really great experience. And if we want to phase out our mpls contracts we can do so over time and was popular initially it got 06:56 up to a certain point but applications really can not really benefit that much again for being optimized. I'm just a single link. So we need to add 07:06 the resiliency in the reliability now. And then against ass and real-time applications like voice and unified Communications are coming into the 07:15

picture. They need to be prioritize. You need to have a really great experience with those they're very subject but you know loss and things like 07:23 that. So it's critical that those applications to work because it patient care really kind of depend on that and it can it can be 07:31 very critical to have those calls be dropped. And then in terms of what we see in today's networks, we 07:41 see a couple of things happening. First of all, we see a lot of organizations today are back hauling all of their traffic whether it's fast or just 07:51

going to their Data Center and so that is causing latency because they're having to send it from the remote site location all the way back to wherever 08:00 the Datacenter is and then out to the internet or to the South's application or going through proxies and and all these kinds of things can 08:09 delay the the packets and caused a poor experience for these remote site users. And then the other thing would be to get to the Cloud 08:18 again, you're back hauling the traffic and then you're you're using a carrier based technology either an email Microsoft express route as your direct 08:28

connect your Google interconnect. They're costly you're still beholden to the carriers in the contracts and all of that, which is really what your 08:38 Avoid someone you're thinking about carriers trying to offer services or even offer asking when you have to ask yourself if you really still want to 08:46 be locked into that carrier type of model. And then the network as a result has really driven cost up 08:55 in terms of it budgets. It lacks resiliency in terms of being able to recover from any kind of degradation or outage 09:04

applications need to be prioritized these days because there's certain critical applications that that can't be slowed down by other types of traffic 09:13 and then in terms of being cloud-ready either for now or for the future when you go through digital transformation, 09:23 So dads going to talk a little bit now about why Northside chose Citrix sd-wan as their solution why we needed it. 09:33 Listen quote on quote. I need you to untangle the sins of the past. We made some less-than-desirable decisions over the course of a 09:41

decade in terms of how we dealt with our when one of the primary ones was all of our Branch sites lack any redundancy whatsoever. That's a bad thing. 09:51 If a branch light fails or early as you can I give you the branch-like pills the doctor packs up goes home close the shop. Now, we've got patients 09:59 that are scheduled for that day. They go home now they're mad because I've got to reschedule come back through all this. So that was one of the key 10:06 drivers we had to win 0ver done it so we had to do something about that. One of our other Major Payne points was we have it said, it's an L2 VPN for 10:12

AT&T call AT&T switch ethernet circus. Bishop one of our pain points. We have a an Eclectic collection of of Transport technology this being one of 10:21 them in this is my biggest pain point that I had to overcome being at Northside. We were moving away from this there's there's issues with AT&T. They 10:30 run a script every week as if you got more than 255 Mac addresses on this transport. We start pruning will now if somebody did something wrong and 10:38 I've got you know, Mac addresses. They're leaking into the win all of a sudden I could potentially prune off. I've had it for enough 30 40% of my 10:47

attached sites because hey AT&T said you got too many Max on here start turning them off regardless of what they are throwing him off. So we would 10:53 have 30 40 sights go down. Nobody knows why we call AT&T. It's a fiasco just getting anything with AT&T if anybody is from AT&T, I'm sorry if you're 11:00 not you understand the pain. We migrated not to a cloud EMR and we're migrating all of these remote practices to VoIP a lot of these remote practices 11:10 again, as I said, we grow through acquisition. So these practices might come in with Vonage pick your Cloud VoIP provider with no control over. We 11:20

just have to deal with it with. Quality issues with that specifically with the emrs. If anybody's in healthcare at you know, that if a doctor 11:29 complaints. Happens really really quick that is one of our pain points as well are EMR, we're migrating to all these these remote practices to iCloud 11:37 base TMR. We have quality issues with that right now based on that l2vpn from AT&T. Obviously the cost of 11:45 will go into this in a little bit as far as what what we saved on this but to hand the doctor a bill and say okay for you to join our Network. It's 11:55

$17,000 up front Just In hardware. And by the way, it's going to be $15,000 a year annual recurring circuit cost and support. They almost checked 12:02 himself in with a stroke. Imma, get back to you know, Mr. Traffic prioritization. I can't prioritize it over this AT&T l2vpn the way it was 12:11 deployed with the equipment was deployed on I can't do anything with it. It's it's hope for the best if somebody complains tell him how the problem 12:21 will solve itself in a couple minutes It'll clear up and it's just not a good solution. It's embarrassing for us. So that's again that was another big 12:30

pain Point going back to that l2vpn and the need a bandwidth again. If you're in healthcare, you understand the whole packs. Imaging radiology Imaging 12:36 nuclear Imaging MRI, and I think it's kind of ironic Valerie put the two slides they were doctors taking images. We have another one that's called 12:45 multi overlapping image slicing or something. It's time to do an MRI of your elbow. It might be 500 slices 500 individual images, but if you get a 12:55 thoracic scan, it could be 2500 images. Each of those images is 50 Megan size. I've got to upload those two are taxes due. And then immediately bring 13:04

them back to the office to the branch office so that the the radiologist or whoever can complete the study that consumes ungodly amounts of bandwidth 13:14 SD. When has really really helped us deal with this problem because of the way we can now leverage both circus simultaneously as opposed to have that 13:22 primary circuit if he's up great, but if he's not a fail over my back up with us when we have the option of saying I've got tax traffic I got the I 13:31 got traffic that is destined for my pack database take that traffic split across both pass and will reassemble the TCP stream on the remote and now 13:39

I've got double the band with double the time or the time now that it take me to get those those those studies back to my data center. 13:48 Back to ya. So we'll cover SV wehen as an overview here since your sister you in and some of the key differentiators. So what does he want and does is 13:58 that takes up to eight physical pass and it creates a logical virtual path and all those links now become active all the bandwidth becomes 14:07 aggregated and we track the conditions across the land in real time so that we can send the packet since we're packet-based 14:17

technology. We can send a package over the best link for that application at that time. We have the ability to identify actually more applications 14:27 than our competitors 4500 applications including sub applications. So if you think of Office 365 we can identify each individual application. 14:36 Then we can look into steering that traffic to wherever it needs to go weather needs to break out locally or go back to the data center or to SAS 14:47 application and then we can do the quality of service or quaas. We can prioritize the traffic over a certain links if needed. We 14:57

have categories for prioritizing our proprietary protocols and were able 15:07 to really do quality of service in both directions. So what are the key points for measuring both directions The Windlass are not measured as 15:16 a as a like an icmp round-trip its measured unidirectional e both directions the Datacenter will communicate with the branch in One Direction and the 15:26 branch back to the data center in that One Direction what that allows the what allows the overlay to do is identify. It's got a full picture of what 15:35

it takes to get from one side to the other and from the other side back. Proximal Phalanx the reality of what happens is you may have a wandling 15:43 wandling k and windlake be wandling a going from the data center to the branch may have 10 millisecond latency, but that return traffic on that same 15:51 way. Like I might be 15 milliseconds. However on when link be returned traffic might be a millisecond so it gives you the ability measuring at 15:58 unidirectional e to say, okay going to the data center me to take this path it coming back. I'm going to go on a different path again. We reassemble 16:07

the stream at the stream reassembles itself. So we have to worry about a separate, you know, asynchronous routing it. It's just it's not it's a it's a 16:14 non-issue pockets in both directions really help save on the usage maximize the usage of your bandwidth 16:21 and make sure that you know, the link conditions in both directions. So it's really helping you on the receiving end as well as the descending and so 16:30 you're not spending too much that does the receiving end can't receive and drops packets. I'm so I want to go into what are 16:39

approaches for citrus SQ in these days and what we can offer so you don't no longer need to back all your traffic. We 16:48 have lots of options for breaking out traffic based on the DPI engine in the identification and classification of applications so we can do direct 16:58 internet break out. Of course, you know, it's it's secure direct internet breakout. So that's one thing you can send it through a swig. We have 17:08 Integrations with those we can do reliable Cloud on ramps. Obviously you've heard about our message real choice and flexibility. We extend that 17:17

through sdn for the cloud so you can choose from Asher Amazon AWS and Google Cloud will it will be in the marketplace there in the 17:26 next couple weeks. So it's really a lot of flexibility and actually a lot of automation that we've brought into being able to connect quickly and 17:36 easily to time to the clouds and and that's important in terms of moving your workloads back and forth or are starting a hybrid Cloud kind 17:45 of scenario. So sad break out. We have a lot of API integration there as well. We have API integration now with Palo Alto globalprotect 17:55

cloud service to connect your IP sax automatically. We have the same thing with zscaler and connecting to there's an answer cuz 18:04 he's killer enforcement. I was in there clouds. So they offer a lot of the next-gen firewall types of functionality that you would need when you are 18:14 thinking about going over the internet. I will talk about some of the integration that we have with Office 365 as well. And then I'm coming soon in 18:23 the next 2, you know few weeks or so. We're going to offer a cloud direct service to this is something that's up very much approve you this is going 18:32

to be about seven tops in the US three in Europe and one in Asia, and you don't actually have Put a nesting ran up there in 18:40 there and not those pops. So we have enough if you have an s u n in the branch, you can then have all of that reliability resiliency quality of 18:50 service Indiana link bonding between the sun in the branch and those pops and then that those pops have a 19:00 service that where there's peering with all of the SAS applications including Office 365. So very quick very reliable. So it's a great way to do 19:09

SATs breakout from the branch. And then in terms of our virtual apps and desktops offering as part of the worst face, 19:19 we actually are the only ones that have the ability to parse into and decrypt Ric a protocol which is the virtual apps and desktops. We do 19:29 the HDX optimization high-definition experience on top of that. But when we person to this protocol and decrypt it, we can get into the individual 19:39 virtual channels that exist within the protocol and then we can actually prioritize that even over a single port so it's called Auto qos or 19:49

autocross. And so that's very important when you're sending trapped the virtual apps and desktops traffic across the land to your branches because you 19:59 can prioritize your real-time traffic over, you know, if I'll transfer a print job. That's something that that is his excellent. It doesn't matter 20:08 where your virtual desktop agents are weather there in the cloud or on-premise. It's the same. Functionality either way 20:17 we're migrating to a cloud-based EMR. It's all citrus-based. I can now from that Branch side over that virtual path 20:25

prioritize. My my my I see a traffic the individual streams and that's been of this going to be a big win for us 20:35 in there where 20:40 everything on the the back end of azure is completely automated from the instantiation into the provisioning to the connection to the MCM. So and then 20:49 you get all the resiliency as well. As you know, the the reliability between the connections from the branches. 20:59 I'm in terms of the Virtual Office 365 experience. So backing up a second. We do have API integration from 21:11

Microsoft for the Office 365 Cloud. They have developed. They continue to grow their at their Office 365 Cloud. They have a huge amount of employees 21:20 or front doors. We now can identify where those are based on the URLs in the endpoint and then we can classify The Office 365 traffic and 21:30 optimize or even allow it through that API integration. So we're able to break out the Office 365 traffic locally and this is very important 21:39 because we support their connectivity principles that they've written and published on the web in terms of reducing latency. So we can now do 21:49

proximate DNS with quad9 and all the things that that can reduce latency in terms of reducing the amount of proxies that you go through or even 21:57 eliminating them because they have the the security that takes Fear of being able to break out locally in terms of teams traffic 22:06 since the media processing is happening on the client side in the branch. We can break that traffic out locally come straight out to the closest 22:15 Office 365 front door again, and then as far as video-on-demand and unified Communications, we can also break that out locally as well. 22:24

So in terms of HTTP, so a lot of different options to steer traffic because we can identify it and because we have the integrated 22:34 firewall were able to do that. And then I just want to take you through a quick down. I want to show you the easy. It is to go through 22:44 the integration with the Citrix managed desktops. It's very automated the Azure functionality happens on the back 22:54 and you have the high of a liability the redundancy the resiliency and it really provides a great experience. It's really just five steps women go 23:04

through very quickly. Can you get that secure connectivity? So the first one we're going to do is right in Citrix managed desktops. You're going 23:13 to out of connection. And then right from there, you can select us you and to give you 23:23 the lowest latency connection best quality fast experience. And really this is the bulk of the work here is doing the 23:33 configuration steps. So setting up your deployment mode Etc. And 23:43 then all of the Azure connectivity as I mentioned happened, so we've got the connectivity there and then we're really just going to monitor the 23:53

house and the connections from that point on. So, how did that automated Azure instantiation is is there really a great way to 24:02 bring that the cloud more and more closer to being a reality for you. So why did 24:12 we choose them? You know we're obviously went through the problems. We are having what are we doing for? This was a big one Superior fail over time 24:22 and full of redundancy. Obviously. I've got redundant Lan connectivity. But the fail over time is key. If you hear if you've done dmvpn, you'll know 24:29

that can't fail over time for a circuit could be 3 seconds 5 Seconds to tell what your timers are brown out issues with the worst because you get a 24:38 brownout. Did you please not going to fail over it was BS not going to fail over. Nothing is fast enough to do to get you through a brownout. I think 24:45 we found it a rule of thumb on a failover was at this point in time. It's two and a half times the latency of the link will cause the traffic to 24:53 migrate from one house to another so for me here in Atlanta all my sites are within I don't know 4 hours of each other from the farthest point of the 25:00

furthest point. My latency is 8 milliseconds From Any Given site to the Head end so I can fail over in the case of a brownout. I fell over in roughly 25:06 20 seconds traffic moves for 20 milliseconds in a second. A real bad day, but no traffic will migrate from one another and it does this 25:16 all day long as late as he changes as Jitter changes as we know we proceed through Brown. I think that this traffic is flows like water around the 25:26 rock in the Stream. It just takes the lowest past the easiest path. That was a key one for us. Now again, as I said, we can now prioritize that I see 25:35

a traffic over Arwen at least from the branch to the data center from the data center to our to our Claudia Mauro. We have two of us set up a little 25:43 bit differently. But from the branch to us using commodity internet service. That's the key. We are a commodity internet service Comcast Business 25:51 class and LTE so we can ensure qos over internet. He will tell you can't do that. Come on over. I'll show you how you can do it. It works. It works 25:58 very very well. And if I get going back to you time to make a routing decision and it's done so on a packet by packet basis. So as as the Box get the 26:08

packet in the date of planets of the cat got to send this witch path. Am I taking a pasty every packet gets that decision made? What's the Lost 26:17 latency and Jitter on a link right now? They put it on this path the next PATH may change, but it over here and it does so all day long and this is 26:25 really flow base right flow Bae story outflow base or session-based. Yeah, and you're getting now you're back to manually pinning it to a pass and I 26:34 don't want to do that and I came back when I started doing this. I was with big large retailer here in Atlanta Publix Supermarkets. We were big on 26:43

deterministic pack of clothes. I want to know from A to B to C. I want to take this path this past this path now like I don't care. I really don't 26:52 care how it goes for me to be as long as it does. So and it does so at with guaranteed experience, which is that what we got now, This is what it used 27:00 to look like still kind of those were still kind of in our Phase 1 deployment right now of RSD when we got about think we're it almost a dozen sites 27:10 that are going to be like next week. But as you can see I said it's an Eclectic mix we've gotten this is just a small slice of it. We got 27:17

point-to-point circuits that terminate Datacenter we've got this wonderful AT&T LTE VPN, which is the bane of my current existence and then we got Dia 27:24 internet circus with ipsec VPN. Every one of these has to be managed individually. It's painful especially on the ipsec side because now we have to 27:33 manage and you know individual ipsec tunnels. Individual proxy ideas just it's very very painful. This is what it is. Now. 27:41 Everybody comes in on the internet. Everybody gets a pair of 75 by 15 Broadband circuits. We do we do acknowledge internally the hay right now as it 27:51

sits in our Phase 1 both of those circus come from Comcast. That's a risk. We've we've been willing to take at this point in time. So the first three 28:00 sites you see this these were the first ones to come up everything post the first three sides now also get an LTE radio in the box and 28:07 that is letting us now leverage. If you're in healthcare you can you can take advantage of firstnet, which is the AT&T band 14 first responder 28:17 Network, which gives you priority over every other cell phone that's around here. So if there's something going on I'm guaranteed bandwidth on Van 14. 28:25

AT&T said that you guys can do this. Perfect put it right there. Now what we're doing is we're taking all of this and we're just saying okay on that 28:33 all to eat that boy PMR traffic. That's it. Nothing else. I don't care about FTP. I don't care about Gmail. I don't that's all secondary get the 28:40 doctor happy and let patients make phone calls. What did we achieve with it? This is the biggest one. When I went to 28:49 my leadership is to head I want to do this because you asked me to fix all this mess that we have. This is why I went home with we're paying right now 28:59

on just got that $17,000 per site just for Hardware. We're now at 6000 maybe 7000 depending on the size of the site and are objects that 29:08 sort it circuits and support on the hardware 12 to $15,000 a year. Go to the doctors say here's your bill every day. It's painful. Our annual 29:17 Optics now for a physician practice is $3,000 and includes two circuits plus support on the hardware on the on the CSUN Appliance and are there to 29:27 switch here in the office now we went from this ridiculous mix of routers layer 3 switches with insanely costly down to $3,000 a year for all of it 29:35

doctors were ecstatic not to mention now that we have complete redundancy from a certain perspective with LTE G&R are dual Comcast. 29:44 This is something we're currently working on. We're waiting on Citrix to drop the code for the glow protect cloud service. The basically here you go 29:54 Point click build the tunnel to offload Our Guest Wireless traffic right now. Our guest Wireless traffic has arrest from the branch to our primary 30:01 data center over our data center connect to our secondary data center where our guest Wireless egress lives. It just adds that much more consumption 30:08

on links that I just really don't want guess traffic on desperately waiting for that feature the day of drops were going live with it. 30:16 As we go back to just a quick note on security. We have security across both the date of plane and the control plane across 30:27 the wire. We have the global encryption me a very strong encryption up to AES 256-bit. We do key rotation every 10 to 15 minutes. We have extended 30:37 packet encryption headers and trailers to prevent spoofing modifications across the wire. We have an integrated stateful firewall. This allows you to 30:47

filter traffic prior to encryption and we have zone-based policies that you can set up for traffic segmentation and consistency. We do have the 30:57 integration as I had mentioned previously with Palo Alto and zscaler actually putting Palo Alto as a VM on our 1100 series Appliance 31:06 up here in the next couple weeks or so. So that'll be a great way. If you're using Palo Alto to continue using that since they manage every everything 31:16 that they offer through one Central Best management tool we have the automated ipsec configuration to the cloud Security 31:25

Services for zscaler Palo Alto. You can see an obviously helps avoid backhauling help consolidate some of the firewall type of 31:35 infrastructure at the data centers and provides a wealth of security features there a real-time inspection. So you really can send all your traffic 31:45 through those aggregation points up in the cloud and then have all that security and all that reporting for you and then in terms of the Su an 31:54 orchestrator, which we just introduced that's now available for the Enterprise as well will be on from later on this year 32:03

in there as well tackle 3DS and then any communication between any Appliance in the estimate orchestrator gets 32:11 encrypted so that that's all secure as well. And back today for so everybody 32:21 needs these pretty critical. What are the failover times and 32:31 how often are the network conditions probed? One of the things we learned was in the absence of traffic. This is just being Payless want to see what 32:41 this thing does in the absence of traffic every 15 seconds. There's a packet going on the wire to prove the path to the data center. Same thing from 32:48

the Datacenter back. What that adds is about 40 kilobits per second. I think is what the math game to for us of just in the in the absence of traffic. 32:56 There's 40 kilobits of traffic that is generated by the appliance to determine what the state of the win is. As soon as you start moving traffic, this 33:04 is where we tried. It was really cool every packet now becomes a probe. So if you're moving a hundred Mega second, give or take roughly with full 33:12 package sizes. You're at about 10,000 packs for every one of those is now a probe on that path multiple paths 10000 packets you now you can see the 33:19

Fidelity to which you can get down to in the granularity of that path in the state of the path. This was a huge horse and then I goes into how quickly 33:29 this can fail over from past the path based on the quality of the path. You have just insanely granular view of each path. This is something I 33:36 think everybody does some people get overly hyped about how we're going to do. This. This is great and you focusing on one vendor and you and you lock 33:46 in on it because I don't know that the bar is a friend or so forth and so on build yourself a matrix. I mean it it seems like it's common knowledge 33:53

with so many people just don't do it. We did it. We went out and we I forgot how many we did. I think it was about six or seven vendors. We put all 34:00 our criteria and they're just highlighted in green which ones were good yellow which one's kind of sorta and then just Graford just didn't make the 34:08 cut. We learned a valuable amount of information from that. If voice important if I mean how many other boys shop? 34:14 I'm not made up. Ask about packet duplication avoid traffic and if it's true pack of duplication by the fall out of the box with we duplicate 34:24

packets over the wire we can take that and do it on any other packet flow that we want to EMR. Are you more traffic when we roll this out to the 34:34 branch? It's a checkbox duplicate to try and duplicate the frames for the EMR traffic. Yeah. Are we consuming bandwidth on both links? Sure. Do I care 34:41 know? Why cuz the doctors not going to complain when one leg goes down or one leg has a brown and I've got packet loss and stuff getting dropped cuz 34:49 I've got a packet in flight on another wire. I don't care van with cheap will just add more if we need it. This is a big one and it can be extended to 34:56

any traffic type which I think is a pretty cool feature. You do with FTP if you want to test if he's important knock yourself out 35:04 but not everybody can do it from a from a vendor prospective either. This is kind of big deal because we build Dynamic 35:11 message between two sites if I've got a practice that needs to talk to another practice. That's the same company are we got again hundred 50 sites of 35:21 those hundred 50 sites a bunch of them have five or six different practices and they're making voice calls between each other. Why do I want a back 35:29

haul that boy traffic from that site back to the day to send it back out that site what this allows me to do is set a threshold if I have traffic 35:37 going from a to the data center to be I can set a threshold. This is if it's more than 80 kilobits per second between these two spend up a direct path 35:44 between the two and let that communication goes a point-to-point instead of a call through the Datacenter once it's done tear it down use it somewhere 35:53 else. So this is a big one and NVR capabilities. If you're if you're so inclined to go down that route, we're looking at it for our guest Wireless to 36:00

get gas. Wireless off our Network beliefs off of our land How you get these order? Look for product maturity? How old is the 36:07 product? Do you know what generation what generation products are you guys now? I know it's 6 6 generation. Maybe 36:17 it's fairly. It's fairly well matured. There's vendors out there that are on product Point 1.2. 36:26 Look for that that's going to get you a long way in the last thing that obviously is he it's got it easy to manage. This is a big one for us. I have a 36:36

junior engineer who he know. He's he's getting good but he is very Junior and he handles this isn't that easy to use you turn it on. I always equated 36:43 to a refrigerator. I tell my wife I love this thing's like a refrigerator you buy a fridge for you set your crisper drawers, you turn it on and you 36:53 walk away every once in awhile you come back and you tweak the humidity on the crisper drawer same thing you fire these things up put them out there 37:00 and go it's it's so stupid easy. We clone sites. We turn up new sites in 5 minutes. It's clone change some IP addresses change the link name send the 37:06

box outside plug-in turn it on low incomes up done is five minute. I don't know anything else that can be done that quickly by a junior engineer. 37:15 What we're doing moving forward. Uq less. This is kind of a little bit of a misnomer here to us as built in the Box by default. There's a 37:25 really what we found is a phenomenal qos set. That's that's default from Citrix. Most people that I talk to you don't ever touch it. They just at its 37:35 greatest does what we need to do. We're going to add to it. We're going to expand a little bit because each of these doctors practices, he sees 37:44

Physicians that they know they go out and I buy their own while I want to use, you know, software alien software be but we have to we hosen the 37:49 Datacenter we want to write individual to US policies for that path. So that that doctors hear that clinicians software can can leverage to kill lice 37:56 and not just get thrown into some arbitrary bolt class or real-time classes because it's Mark differently. Ma scoring this is kind of a big one. Most 38:05 people use it for VoIP. Most people reference it as a voice thing, right? What's your mas? Core4voip mean opinion scoring? How well is a boy 38:15

performing over the win? You can use it for any application. It's just to find your application. Check the box. Login Jesse Wayne Center and start 38:21 looking at how my applications behaving over the way in now, so I'm finishing calls up until this is not working, right? Okay. Well I can log first 38:29 thing I do is okay. We log into an Center. Show me the mosque or for your application. Where is it at? 3.74? Okay, when is not a problem that 38:37 immediately ruled out the way and the problem is you don't like my side or theirs but it's not me. Let's go talk to somebody else who can help figure 38:44

this out. So this is this is something we are in the process of rolling it out again, as I said to Palo Alto globalprotect cloud service for guest 38:51 Wi-Fi or by calling it. We want to get rid of this back hall in it, but we want to keep it within our Palo Alto Panorama installation. So we have a 38:59 single-pane-of-glass to manage all of our content filtering and everything like that. So we're looking at that. That'll be rolled out as soon as soon 39:06 as you guys drop the code pack of duplication with EMR that said single checkbox that's coming up very shortly. We're going to start looking at but 39:14

most of our hospitals are most of our sites within 4 hours. So everything shipped one location. We generally can figure out there send it out, but we 39:23 may start we may start looking at this just to see if it brings us any benefit to just drop shipping and having a vendor plug it in for us. One of the 39:30 things we did do I intentionally we deployed in a single data center. It's at a collocated facility where I don't have to worry about bandwidth. It's 39:39 probably one of the fourth largest data centers United States here in Atlanta, but not to worry about man with power who makes sense put it there. I'm 39:46

not going to have a failure there in if I have a failure there. It's it's me Facebook Google Twitter. Everybody else is going to have that failure. So 39:53 let's put it there. We don't worry about it. So one of the things we have on the radar for 20/20 is hey, let's let's build that second head end in in 39:59 our primary or secondary data center. And then again to continue transition, let's get off this l2vpn. We have to get off this thing as 40:06 circuits term. We're just immediately cutting it spitting up with the pair a Comcast orcas and Anna Sim in the modem. And boom. There we go. That's 40:16

that's really where we that's where we're looking. And then just a couple of notes Here. We have a text in one area. So 40:23 come check that out. It's actually really cool. There's some really great videos are about 15 minutes long are some great documents there and there's 40:33 the mess you and as well as work space and we're expanding and growing that. Before you leave there is a couple other sessions still coming up for 40:40 Sun. So if you want to make note of this we will have the surveys available via email and in the mobile app start Thursday at 9 40:48

a.m. And you can watch sessions on demand as well starting Wednesday today, and then you can start downloading sessions on June 3rd seaweed going to 40:58 the Citrix Synergy webpage and login. And then if you can just do the survey in the mobile application for us, please. 41:07 So thank you very much and please come up if you have any questions. 41:17

Cackle comments for the website

Buy this talk

Access to the talk “Citrix Synergy TV - SYN146 - #RealWorldExample: SD-WAN eliminates the consequences of network...”
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 - SYN146 - #RealWorldExample: SD-WAN eliminates the consequences of network...”
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