Cloud-Ready Data Center Solutions for Private and Public Environments | Juniper Global Summit

00:07 Hi, I' m Mike Bushong,

00:08 Vice President and leader of Juniper's Cloud-Ready Data Center business.

00:12 And let me add my welcome to Juniper's 2021 summit.

00:16 Over the next few minutes,

00:17 I want to provide a view into what we've been working on in the data center.

00:20 And following that, we want to give you an up-close

00:22 look into one of the real highlights for our portfolio.

00:25 In that demo, we hope to make our vision more concrete

00:28 so you can see what we can do for you today.

00:32 Before I get into the 'what', though,

00:34 I want to start with the 'why' that underpins everything we do.

00:38 As you've heard a few times today, as a company,

00:40 we're relentlessly pursuing the big idea of delivering an Experience-First portfolio.

00:45 And this extends, as you probably expect,

00:47 to our data center offerings as well.

00:50 So what does it mean to be Experience-First

00:54 In a sentence, as you heard Rami say, it's the acknowledgement

00:57 that we need to move beyond building better networks

01:00 to making networking better.

01:02 Networks are nouns.

01:04 They're defined by the hardware, the software,

01:06 the protocols, the technology.

01:09 Networking is a verb.

01:11 It's defined by the people, the process, the tools.

01:14 If we want to move data center networking forward,

01:17 we need to use the noun to transform the verb.

01:20 We need to move past speeds and feeds.

01:23 We need to untether ourselves from some of our legacy roots.

01:27 Where the past was dominated by the command line

01:29 and the vendor certifications required to demonstrate proficiency, the future,

01:34 and really, the present,

01:36 is about leveraging software to change how we think about operations.

01:41 But even how we think about software has to change.

01:45 The goal isn't to recast the command line in the cloud.

01:48 Wrapping up the same basic way of thinking

01:51 and calling it software-defined or as-a-service isn't helping anyone.

01:55 Superficial changes like this are an empty gesture,

01:58 and they' re certainly not enough.

02:01 If we get through a generational change in operations,

02:04 that's what cloud is by the way,

02:06 and all we do is change the point of interaction from the device to a controller,

02:11 we will have collectively failed.

02:14 Operations is so much more.

02:18 And it doesn't begin on Day 2.

02:20 At Juniper, we believe operations starts with planning,

02:23 continues through deployment,

02:25 and of course includes the continuous monitoring and upkeep of the data center.

02:30 It's about getting things right from the outset and making sure they stay right.

02:36 This vision is our guiding principle,

02:39 and it drives everything we've been doing in the data center

02:42 since the last time we all came together.

02:44 Today, we want to give you a peek into what we've been up to,

02:47 we want to show you what operations can look like,

02:50 and we want to leave you with a few clues about what's next.

02:53 Everything we do to evolve data center operations

02:56 is built on some key foundational elements.

02:59 Our data center portfolio is anchored by our QFX and PTX,

03:03 and soon, ACX switches.

03:05 We have built out cost-optimized 25 and 100Gig platforms,

03:09 and we remain absolutely committed

03:11 to helping organizations navigate their own transitions.

03:15 In the past year, we expanded our hardware portfolio

03:18 with the industry's first 400Gig Trident 4 based platform.

03:23 And it all runs Junos,

03:25 which means you get the familiar One Junos experience

03:28 featuring the world's most advanced routing stack.

03:31 But if operations is the focus,

03:34 the foundation needs to include more than just hardware.

03:37 The most basic premise of automation is 'see something, do something'.

03:42 And that means that telemetry plays a leading role.

03:45 In Junos, we've always been the leaders in all things telemetry,

03:49 from programmatic access to streaming telemetry.

03:52 And as the world moves to in-band telemetry, we'll be there as well.

03:55 And that brings me to our primary focus: data center operations.

04:00 When it comes to operations, there are different approaches.

04:03 For companies where the infrastructure is part of the product,

04:06 think cloud providers and SaaS companies,

04:09 the goal is to eke out every efficiency possible,

04:13 leveraging commercial and open-source software supported

04:16 by a DevOps or NRE type framework.

04:19 Here, it's not enough to deliver a product.

04:22 We need to provide a practice that combines the underlying technologies

04:27 with a lightweight support and services surround

04:29 to help companies contextualize their operational practices.

04:34 Not every company can be Amazon or Netflix, but with a little bit of help,

04:38 we believe we can take the first several steps with you.

04:41 If you've got automation ambitions but haven't had the time or space

04:44 to make as much progress as you'd like, we need to talk.

04:49 For other companies, where IT is an enabler,

04:51 the focus is on simplifying operations.

04:55 These companies trade in agility and reliability.

04:58 And this is where our really big news plugs in.

05:01 If operations is important, then what better move than

05:05 to go out and acquire the industry's best.

05:07 Earlier this year, we brought Apstra into the mix.

05:11 They're the pioneers of real change in this space.

05:13 Apstra literally coined the term "Intent-Based Networking".

05:18 They are built around the idea that you ought to specify what you want,

05:21 not how it's implemented.

05:23 And the best part?

05:24 Apstra does this in a way that allows multi-vendor deployments,

05:27 legitimately allowing you to run our solutions alongside our competitors.

05:32 No one else is even trying to do this.

05:36 And of course, networking isn't the only piece of the stack.

05:38 We've built integrations into VMware,

05:41 we have the single best management solution for SONiC with Apstra,

05:44 we've been driving cloud-native networking with Kubernetes using Contrail as an SDN,

05:49 and we have technology that plays into the broader Multicloud ecosystem.

05:54 So that' s a fairly condensed view of our strategy.

05:56 We' re excited about what we' ve been up to.

05:59 Our customers have been telling us that we' re leading where they want to go.

06:03 And you know who else seems pretty bullish on where we' re headed?

06:06 The analysts.

06:07 This past year, Juniper was named a leader in Gartner' s Magic Quadrant for Data Center

06:12 and Cloud Networking for the third straight year.

06:15 And with Apstra as one of Gartner' s Cool Vendors,

06:17 we think the combination of Juniper and Apstra is even stronger going forward.

06:23 So clearly, while we' ve accomplished a lot,

06:25 the highlight of our year of disruption is bringing Apstra into the portfolio.

06:29 So let's spend a bit more time there,

06:31 and there's no one better to talk about how Apstra

06:32 will change your data center experience than Mansour Karam,

06:36 Apstra' s former CEO and now my partner for all things data center operations.

06:41 Hello, I am Mansour Karam, vice president of Products

06:44 and former co-founder and CEO of Apstra.

06:48 Running the data center can be challenging.

06:51 You may be struggling with the tradeoff between agility on one hand, that is,

06:56 doing something fast to meet business demand, and reliability, on the other hand,

07:03 that is, doing it right, to avoid mistakes.

07:07 You may also be concerned about constrained vendor choice, or lock-in.

07:15 Indeed, the overwhelming majority of management solutions

07:19 on the market today only work with one vendor.

07:24 And adding to this, you may need to overcome critical sourcing and skill shortages.

07:31 In a recent Gartner study,

07:34 61% reported that a lack of staff expertise

07:39 and skills were a challenge to their organization.

07:45 Against these set of challenges,

07:47 what is Juniper bringing in the Apstra system that is different?

07:53 1 - Better Day 2 Operations

07:56 Unlike other solutions for Day 2,

07:59 the Apstra system starts with the network design.

08:03 It captures in business terms, in a single source of truth,

08:08 what is expected, what we call intent.

08:11 Then, it consistently checks against the expected state, through rich telemetry,

08:18 finding what matters, and proactively alerting on just the real issues.

08:25 2 - A Turnkey and Unified system

08:30 As a software-only solution,

08:32 Apstra is quick to install, and then it automates all aspects of your network:

08:38 the design, the deployment, and the operation.

08:43 In this way, Apstra is one unified system for both architects and operators.

08:51 And, 3 - An open and Multi-vendor approach.

08:56 Apstra can help you manage your network in a single or multi-vendor environment,

09:02 to include not only Juniper, but also SONiC, VMware, Arista, and Cisco amongst others.

09:12 Let' s take a look at what the Apstra solution is.

09:16 The Apstra-Intent Based System begins with the design functions of the Architect.

09:23 With Apstra, you no longer need to worry about

09:26 specific VLANs or other arcane configurations.

09:30 In one unified solution, you simply describe what you want, for example, reachability,

09:37 security, quality of experience, or compliance;

09:41 and the software takes care of all of the details of the HOW,

09:46 with Apstra' s pre-built templates automating the design of your network;

09:51 and blueprints that fully automate the processes of building and deploying the network.

09:59 Then, the Apstra system extends to network operators

10:04 with continuous automation, analytics, and assurance.

10:10 It' s there to alert operators when brownouts and operational deviations occur,

10:17 along with conveying why the alerts happened, through root cause identification.

10:23 Specific functions for controlling and managing change close the loop to

10:28 deliver the fully Assured and Automated Data Center network.

10:33 No one else in the industry has this: a system for describing the way

10:38 that the network fabric should behave,

10:40 and software that then automatically translates those requirements

10:44 to configure the fabric accordingly,

10:47 assures it is so,

10:49 and also let you proactively know when and why an issue may be brewing.

10:55 Juniper Apstra lets you set your operational model around

10:59 what YOU need in YOUR network,

11:02 not by a vendor' s dictated way.

11:06 Apstra lets you readily adopt automation,

11:09 throughout the entire lifecycle, to automate every day.

11:13 Starting with DAY 0, you use Apstra for design, pre-validation,

11:19 and installation, with all the details covered.

11:23 For example, Apstra will render your cabling plan.

11:27 Then, you continue with your DAY 1 deployment,

11:31 with support for zero-touch provisioning

11:33 and pre-built blueprints for the most common network setups

11:38 and a way to easily tune these for your environment.

11:43 And completing the lifecycle -

11:45 you use Apstra as a powerful tool for DAY 2 and beyond operations,

11:50 with visualization, Analytics, Insights,

11:55 Validation, Compliance, Root Cause Identification,

11:59 Troubleshooting, Optimization, and a structured means for change.

12:06 You've heard enough from me.

12:08 Let' s take a closer look at how it all actually works!

12:12 I' d like to introduce you to Scott Sneddon

12:15 and Jeff Tantsura for the highlight of our session, our Demo Challenge.

12:21 -Thanks, Mansour. My name is Scott Sneddon.

12:24 I' m joined today by my good friend, Jeff Tantsura.

12:26 We' re on the data center team here at Juniper Networks.

12:29 Jeff, I' ve been working with the team for a little while now.

12:32 They' re working on moving away from their vintage network

12:35 hardware and some of those legacy designs that they have in place.

12:38 They really like the idea of EVPN-VXLAN, IP fabric, all of that stuff.

12:43 We want to get going on the design.

12:45 We want to get going on getting things implemented.

12:48 We' ve got a whiteboard diagram, but we need to move to the next step.

12:51 Let' s say they want to build a leaf-spine two-pod data center.

12:56 Getting into the detail of all that can oftentimes take days, weeks.

12:59 There' s a lot of stuff that needs to be captured.

13:01 Can the Apstra intent-based system help us with that?

13:05 -Hey, Scott, great to see you, and with pleasure.

13:08 I' m going to guide you through design steps needed to build

13:11 data center according to business logic you expressed.

13:15 When we start building data center, we always start with design phase.

13:18 I' m going to show you a number of pre-built templates

13:21 we are going to use to build the data center.

13:23 We start with ref templates.

13:25 I' ve pre-built two, one of them with two leaves

13:29 and four servers with no external connectivity.

13:32 Another one that' s very similar but with external connectivity

13:37 that is validated during the template building phase.

13:44 Now, we are going to apply the ref template we have

13:48 just built into template of the data center.

13:52 What you' re going to see now is how your data

13:54 center is going to look when you build it.

13:57 Exactly as you said and according to your intent and business logic,

14:01 we are going to have couple of super spine

14:04 and two pods with two leaves each.

14:09 Now, we are going to apply template to blueprint.

14:14 Blueprint we are building,

14:16 we are going to call it Scott because I really like Scott.

14:19 We are going to match the template we' ve pre-created and attach it to blueprint.

14:26 What you see here is exactly your network-ready model.

14:32 All the parameters that needed to be there, we already know them.

14:36 It' s going to be a couple of leaves.

14:38 Now, we are going to see the configuration on the leaf is completely empty.

14:42 We have no VLANs configured, we have no routing instance.

14:47 The only routing instance you see is for management

14:50 or really next-hop to the management interface.

14:52 Since we have validated on the device itself that it' s empty,

14:57 now we can go on and configure services.

15:02 If you want to see full video, please go and watch it on YouTube.

15:06 What we have configured here is security zone, that is, layer-2 segmentation

15:11 called Scott as well as virtual network that' s layer-2 segmentation

15:15 that, guess what, we called again Scott.

15:17 Now, we are going to commit the changes.

15:20 When you use ASN configuration, you' re always stage four so you can review

15:24 all the changes and you never push configuration without reviewing it.

15:29 Given we have reviewed it and it' s what we want,

15:31 we are going to commit it,

15:33 and we are going to call it Scott again.

15:36 At this point in time, we take logical data models, random transformation,

15:42 generate configuration that is going to be pushed to the devices.

15:47 Now it' s done, network is fully configured, has all the parameters,

15:52 all the routing semantics, everything we want it to be.

15:55 We see that everything is green, meaning, system reports

15:58 that configurations that had been pushed, if valid, has been applied,

16:05 and derived states are correct. Everything is green.

16:07 This is how we want to check our network.

16:09 Let' s check the devices. Now, we are looking for routing tables.

16:12 Guess what, we see tables called Scott, we see Scott EVPN tables

16:17 that has all the EVPN route for the routing instance Scott

16:21 as well as inet.0 and inet.6 table that have

16:26 all the route that are related here.

16:31 Most importantly, the BGP-EVPN table has been instantiated

16:34 where we keep all the routes that came from other EVPN routers.

16:40 Now, we are looking at routing instances.

16:42 You could see there' s new routing intances called Scott that test

16:45 all the parameters that required for EVPN-VXLAN to work.

16:51 The VNI ID is 10,000 as well as route-target.

16:55 It' s provided by the system. We know exactly what it shouldn' t be.

16:59 We don' t use any automatic options and it' s all predictable.

17:02 As you can see, we' ve configured virtual networks, security zones, all done,

17:08 and device on this date validates all the configurations working correctly.

17:13 All the configuration in place. Back to you.

17:18 -Wow. That was quick. I barely had time to get a cup of coffee.

17:24 This is clearly saving a ton of time with getting things designed and deployed.

17:30 A lot of customers want to automate, but they maybe

17:34 don' t have the time to spend on the learning curve to get things automated.

17:38 They also might not have expertise in BGP, and you' ve shown us

17:42 that the Apstra intent-based system configured all that for them.

17:46 That' s a really powerful tool.

17:50 Now the app team just called and they need to deploy

17:54 that app that they' ve been developing tonight.

17:57 I guess the marketing team is going to run a big social media

18:00 push and they' re expecting a ton of traffic to come hit that app.

18:04 As with everything, it' s a fire drill and we' ve

18:07 got to get that app pushed out and ready to go.

18:09 That means a bunch of stuff has to happen on the network.

18:12 You created the EVPN contacts and the VLANs, and all

18:15 this stuff to accommodate that app, but hey, guess what?

18:18 That app needs to run in both data centers.

18:20 Now we' ve got to figure out the data center interconnect portion of that.

18:23 Can the Apstra intent-based system help with that?

18:26 -With pleasure, and this is where we shine.

18:29 What we are going to do now we are going to create EVPN gateway that will

18:33 interconnect to data centers together and create a common network.

18:38 We' ll go to workflow called create EVPN gateway, call it, again, Scott.

18:43 Provide its IP address, its ASN number,

18:48 and let' s say it' s going to be 10 hops away, so TTL of 10.

18:53 We' ll keep the rest of parameters.

18:57 What' s important now since EVPN gateway is a logical function,

19:02 system allows us to choose how resilient we want to be.

19:05 It could go on single device, or probably you should do at least

19:09 two for resiliency if one goes into maintenance, or just break down.

19:14 You could see now that we are going to choose two leaves

19:17 in different ports for resiliency to provide this function.

19:22 I' m going to choose two different leaves where EVPN

19:27 gateway function will be instantiated and commit a configuration.

19:33 As always, configuration is staged but not committed.

19:37 In order to make it work, we are going to commit

19:40 it and this is when it' s going to take really place.

19:46 Let' s validate first what we see, Scott, IP address, ASN,

19:52 TTL, and default parameters for hold and keep-alive.

19:57 It' s instantiated in two leaves.

19:59 From configuration perspective, everything is right and correct.

20:04 We can go and commit the configuration.

20:17 Scott, we are done.

20:20 You' ve got fully working EVPN connectivity between two different data centers.

20:26 [music]

20:30 -Oh, back to me already. Wow.

20:33 I thought I' d have time to lay down a track

20:35 or two while I' m here in my home studio.

20:40 We can get everything going really quickly.

20:43 We' ve got the data center interconnect built, that application' s ready to roll.

20:48 Now, here' s a scenario that never happens in reality.

20:51 The app team called and there' s a bug,

20:53 and they need to roll their application back.

20:56 I guess we' re going to postpone that marketing

20:57 launch and the app team needs to go debug their code.

21:01 We' ve got to pull that network configuration out.

21:05 We' re going to roll back that data center interconnect,

21:07 we don' t want to leave that stuff sitting there while they' re troubleshooting their app.

21:11 I know in Junos, I could log into the CLI and roll back the config.

21:15 All of us that have been using Junos for a while, love that feature.

21:20 In the old model, I bet I could get that done in 30 minutes,

21:24 log into each CLI and hit rollback on each of those switches,

21:27 and validate that everything' s done,

21:29 but I bet the Apstra intent-based system can beat that.

21:32 Jeff, do you think you can get that done in a minute?

21:38 -I' ll take a challenge. If you lose, you owe me a beer.

21:42 -You got it.

21:42 -We are going to use functionality we call

21:45 Time Voyager that allows you to move in time.

21:50 Anytime you commit, we create a snapshot of whole network,

21:55 all its states all configurations, and save it.

21:58 If you need to roll back the previous state or one before, you can

22:01 easily do it by rolling back your system exactly to this state.

22:07 As you can see here, last state known state is gateway.

22:11 As long as we don' t configure the gateway.

22:13 What we are going to do now is to go back to Scott and hit the timer.

22:20 Go.

22:25 Staged.

22:29 We check what it is. We have removed EVPN gateway.

22:31 We are going to commit.

22:48 Done. You owe me a beer, Scott.

22:51 Now, we are also going to prove

22:53 that configuration had been removed correctly.

22:56 We are going to take a look at our network

22:58 and see whether DCI gateway has disappeared.

23:02 Let' s take a look at virtual network.

23:05 You could see EVPN gateway directory is empty.

23:08 If we look at complete network, we will see that EVPN gateway has gone.

23:15 Validation of our intent is that we are back to previous state.

23:21 Back to you, Scott.

23:22 -I' ll gladly buy you that beer anytime, Jeff.

23:26 It' s really a big time saver, this Apstra intent-based system,

23:31 introducing a lot of automation to do a lot

23:33 of the tasks that need to happen with an operator.

23:36 What about troubleshooting? What about monitoring?

23:38 What about the day to and beyond sorts of things?

23:41 What can we do to reduce the meantime to innocence?

23:46 -This is very important point, and thank you for bringing it up.

23:49 On system side, system always validate the configuration the devices

23:53 run is in sync with the configuration as produced by the system.

23:59 Anytime there' s any change, system will make

24:01 sure to notify you there' s configuration.

24:04 If someone actually goes and touches the system to CLI,

24:07 it' s an anomaly from system perspective.

24:10 We will notify you immediately. Let' s try to break something.

24:13 It' s called breaking stuff. Let' s actually break the system for CLI.

24:19 What we are going to be doing, now, we are going back

24:22 to all great Junos CLI to disable an interface.

24:31 Let' s take a look at what we need to do here.

24:38 We are going to disable interface that leads leaf

24:42 to the spine, that connects leaf to the spine.

24:44 Before we do so, we need to validate that' s actually the correct interface.

24:48 Let' s go back to the system, take a look at leaf one

24:51 and see which link connects leaf to the spine.

24:54 As you can see, it' s actually 000,

24:57 and this is actually the interface we are going to shut down.

25:00 Let' s go back into CLI and disable it.

25:04 Let' s commit it now.

25:11 Our fabric is officially broken.

25:14 You see red anomalies, and obviously there.

25:18 What you also see, there are dynamic anomalies

25:21 coming from BGP, coming from your dynamic state.

25:24 Let' s take a look at what happened to configuration.

25:27 As I said, we immediately notify you that there is a difference between

25:31 expected configuration as driven by the system and its data model

25:35 and actual configuration assumes the device.

25:38 This is exactly what we call configuration sleep.

25:42 If we take a look at actual, we will see the delta

25:47 which is exactly the change to state of XE interface we' ve disabled.

25:57 The system has notified you about all the inconsistencies

26:02 and allows you to fix them in number of ways.

26:06 You could go and manually roll back.

26:09 You could use Time Voyager or you could use function we call accept new configuration,

26:14 and new configuration will actually be accepted.

26:17 It' s very useful for troubleshooting purposes

26:20 when you need temporary change configuration.

26:23 We help operators a day to validate all

26:27 the changes and configurations that were intended.

26:30 We will notify if there are any changes to dynamic

26:33 state BGP sessions, BGP route, next hop.

26:36 Anything that' s related to routing is done by the system.

26:39 Back to you, Scott. Did you like it?

26:41 - That' s pretty slick, Jeff.

26:44 I could really see this being a useful tool, a cool

26:48 tool for operators, great automation capabilities,

26:53 really being able to abstract a lot of the complexity

26:56 of EVPN-VXLAN into simple templates and blueprints.

27:00 Then just think of the anomaly detection.

27:04 There' s a whole lot more there with analytics that we can unpack as well.

27:08 For our viewers, I encourage you to check out some of the YouTube

27:11 videos and the longer demos that we' ve got available to you to go deeper

27:16 into the intent-based analytics capabilities and a lot of the Day 2

27:20 operational aspects of the Apstra intent-based system.

27:24 With that, Jeff, I want to say thanks, really appreciate the demo today.

27:30 To our viewers, thanks for your time. Mansour, back to you.

27:35 Thank you Scott and Jeff.

27:37 Wasn' t that cool?

27:38 In fact, after our customers have used Apstra for a while,

27:42 they tell us the biggest difference is in their Day 2 Operations.

27:48 It' s always about the time saved in Day 2 and how that translates to speedy changes,

27:55 issues averted, rapid resolution, and fast time to innocence.

28:01 In this way, Apstra delivers, in a fully automated way,

28:06 across the spectrum of Day 2 needs with: Visibility

28:12 and Insights Incident Management Root Cause Identification

28:16 Change Control Change Management

28:20 Compliance and Audit AND Maintenance and Updates

28:26 A customer story helps to sum up the value delivered by Juniper Apstra:

28:32 A Fortune 50 Energy company was in the process of upgrading their infrastructure

28:38 and knew that starting with proper automation was going to be critical

28:44 to meet their infrastructure transformation agility and reliability goals.

28:49 They were influenced by the hyper scalers' approach

28:53 of fully automating their data center operations at the management plane.

28:59 However, same as many other Fortune 50 enterprises,

29:04 they were aware that they didn' t have an army of developers.

29:09 They needed a turnkey solution that did most of the heavy lifting for them.

29:16 They also had enterprise specific requirements,

29:19 such as support for EVPN/VXLAN, which hyperscalers don' t have.

29:25 The Juniper Apstra approach checked all the boxes.

29:29 It was multi-vendor, decoupling the management plane from the hardware

29:35 and enabling them to separate the software and hardware life cycles.

29:41 This is unique to Juniper' s Apstra solution.

29:46 They also found Apstra to be powerful, intuitive, and easy to use.

29:52 They explored replicating Apstra with Ansible,

29:55 and realized that it was impossible to do so because of their Day 2 needs.

30:00 By using Apstra, they now have the ability to deploy fabrics

30:05 and make Day 2 changes using a couple of clicks,

30:09 at a fraction of the time that it used to take them,

30:13 for example, when they had to upgrade their interfaces from 40GigE to 100GigE.

30:19 They value that Apstra provides guardrails,

30:23 which means that network engineers can' t make ad-hoc changes.

30:29 Changes need to follow a certain pattern specified in the Apstra blueprint.

30:35 This is key to infrastructure reliability and predictability.

30:40 All changes are self-documented and can be audited,

30:44 which also helps with compliance.

30:47 Last but not least, closed loop validation ensures changes

30:53 have the intended effect.

30:57 So, what does it mean for you?

31:00 If you' re like me,

31:01 and you don' t want to miss out being with your family and kids as often as you can,

31:06 for me it' s my three young daughters that are the treasure of my life.

31:11 If you' ve been late to a soccer game,

31:14 delayed by a lengthy troubleshooting session.

31:18 Or, if you' ve spent so many long hours typing the same,

31:23 repetitive commands into a keyboard,

31:25 device-by-device, that you were late to dinner with your family.

31:31 Or, if you' ve had to rely on hope for the network to stay up,

31:36 while waiting for an early morning change window to apply a patch.

31:41 Or, if you' ve been waiting and waiting on the promises

31:47 of another vendor to get your automation implemented right.

31:51 Then the Apstra solution can help you.

31:55 It' s not too good to be true; and it' s available NOW.

32:00 And, it' s back to you Mike to wrap our session.

32:03 Thanks Mansour.

32:04 And thanks to Scott and Jeff for the demos.

32:07 I'll be honest.

32:08 I haven't had this much fun at work since, well, since ever.

32:12 It's incredible what's been going on with our data center portfolio.

32:15 Internally, there's just this palpable energy.

32:19 But even better than that, there's a persistent buzz with you,

32:23 an excitement in our customers and partners that's really pretty impressive to see.

32:28 Yahoo Japan uses Apstra to operate in one standard way

32:31 across three different vendors.

32:33 They told us that Apstra allows them to complete network tasks

32:37 that used to take days in tens of minutes.

32:40 For T-Systems, one of the largest data center operators in Europe,

32:44 Juniper and Apstra help to introduce new services,

32:47 quickly scale to new demand, and deliver assured experiences.

32:51 Here, the multi-vendor capabilities are critical

32:54 to extending operational advantages beyond just the Juniper infrastructure.

32:59 At Aston Martin, with Juniper solutions,

33:02 they' re looking to use automation techniques such as Infrastructure as Code.

33:06 To Aston Martin, that means defining

33:08 what they want, and monitoring what' s being delivered.

33:10 We've been fortunate enough to sit down with many of these customers

33:13 to talk with their leaders about how they've met the growth

33:16 and agility requirements of the business

33:18 without sacrificing reliability, performance, or security.

33:22 After this session, make sure you check out these customer conversations.

33:25 If anybody wants to see more of these solutions

33:28 or get their hands on them, there's two things you can do.

33:31 And they're both free.

33:32 First, attend our upcoming data center boot camp in May, showcasing Apstra.

33:37 During that half-day session,

33:38 we'll spend most of our time in demo and looking at real examples.

33:42 Our focus will be introductory,

33:43 and so don' t worry if all you know is not much more

33:46 than what you' ve heard in the last half hour.

33:48 Just click the button on this page to register for the bootcamp.

33:51 And the vLabs folks have got a couple of sandboxes up

33:54 and running, which are accessible for people to get into.

33:56 Head over to vlabs. juniper. net.

33:59 If you don' t have an existing Juniper account,

34:01 you' ll need to take that step first.

34:03 Or, reach out to your Juniper partner or rep right now.

34:06 Let me just close with a few parting thoughts.

34:09 The data center of tomorrow shouldn't be the same as the data center from a year ago.

34:14 Or 10 years ago for some organizations.

34:16 The technology exists to do it better.

34:20 For organizations, plotting your path through change is hard.

34:25 Choosing the right technologies and partners can feel risky.

34:29 The key is making strategic choices alongside people you trust,

34:34 and moving forward together.

34:37 For individuals, there' s this undercurrent in our industry

34:41 that these technologies are here to replace you.

34:44 That couldn' t be further from the truth.

34:47 But we can' t move forward if we spend so much effort just to keep from falling back.

34:53 We, Juniper, are here not just to help your organization,

34:58 we' re here to help you, because ultimately, change is a lot

35:01 more than just adopting new technology.

35:04 And as hard as change is, it's even harder

35:07 when all you see is hype and messages intended to confuse.

35:12 What we' re trying to do here is cut through the noise and just get real.

35:17 Everything we've talked about today is real.

35:20 The tech. The examples. The customers. And the demos.

35:25 And all of this real,

35:27 that' s what' s going to deliver both better networks and better networking.

35:33 It' s time to evolve data center operations.

35:36 It' s time to put experience first.