Video: How Cloudflare can coexist alongside Zscaler to fast-track your remote access project | Duration: 3216s | Summary: How Cloudflare can coexist alongside Zscaler to fast-track your remote access project | Chapters: Modernizing Remote Access (35.95s), Modernization Phases Overview (315.76498s), ZPA Deployment Challenges (439.95s), Clientless ZTNA Adoption (597.975s), Clientless ZTNA Deployment (924.35504s), CloudFlare Benefits Overview (1402.535s), Zero Trust Maturity (1625.5349s), Secure Gateway Policies (2261.07s), Cloud Network Transformation (2656.9802s), Conclusion and Recommendations (2872.385s)
Transcript for "How Cloudflare can coexist alongside Zscaler to fast-track your remote access project":
Hey, everyone. Thanks for joining today's webinar. Today, we're gonna be talking about ways to modernize remote access and specifically highlight opportunities where, even if you are an existing Zscaler customer, you can take a dual vendor approach, and maybe that's your best strategy to make progress even faster with your modernization projects. So we're gonna provide our recommended strategy on how to accelerate your SaaS adoption or secure access service edge adoption. And I'm James Chang. I've been with Cloudflare's SaaS and Workspace security portfolio for five years in a product marketing capacity. I'm joined by other subject matter experts on this topic. I'll let them introduce themselves starting with Marcel. Hi. My name is Marcel Hoffman, and I work in technical marketing. I'm actually celebrating my two year anniversary with Cloudflare today. In my past life, I built and managed large scale security operation centers for global companies such as Sony HPE. And I would like to thank you, James, for inviting me to give some industry perspective on this important topic. Thanks, Marcel. And Chris. Hi. I'm Chris Lopez. I'm the technical marketing engineer, for CloudFlare Zero Trust, and I'm also associated, with our with our competitive suite. So I'll be giving some insight into how exactly Zscaler, can interoperate with CloudFlare in a way that is harmonious. Excellent. Thanks everyone for controlling. So let's just start with some level setting. This is a typical journey that we partner with customers on. It's about moving from the picture on the left to the right. So on the left is your classic legacy castle and moat model with a lots of on prem tooling. And instead, organizations are trying to shift security and network productivity to a cloud platform and control plane. And this unified cloud native approach is also known as SASE or Secure Access Service Edge in that architecture. And in the process, organizations want to enforce zero trust best practices based on default deny and least privilege security policies. Now at this point, most organizations agree on the merits of this type of architectural transformation. But undertaking that journey is a pretty transformative effort, and we recognize those types of projects are inherently hard. They may not always go so smoothly even if you're going with an established vendor like a Zscaler. And so many times, you wind up stuck just like our friend Winnie the Pooh here. And in fact, you may remember that off-site stat from McKinsey that 70% of digital transformation projects fail to meet their stated goals. And so, you know, Marcel, I was hoping, based on your experiences, you could share some perspective on why you see organizations maybe get stuck with these types of large scale SaaS transformation projects. Yes. Thank you, James. So in our experience, there are, like, three very common reasons why these SaaS projects sort of get stuck. Right? Number one, especially when we talk about zTNA, it's a full user experience. Right? Many customers started their SASE journey with a secure web gateway. Right? They in the past, we were all running on prem proxies. It was easy to move that into the cloud, and there were many, many benefits to that. The truth of the matter is that zTNA has very different performance requirements. Right? ZTNA has to be fast, and it has to be reliable because the applications that we still run on prem are probably the most important applications that you have. That's why they're still there. And if anything goes wrong with, your z t and a solution, you will get a lot of tickets from your employees. So we've seen also that many employees then revert back to workarounds, turning off their their agents or try to go around your security controls to get their work done. So that was the number one experience we we heard. The other one is, many, many vendors nowadays try to sell you, a platform. Right? They throw in ZTNA for free, but also give you very aggressive on ramps, which means you have to roll out ZTNA very quickly. Otherwise, you end up with licenses that are just sitting on a shelf. You paid for it. You can't use it, and you just revert back to what works and still run your VPN in parallel. The other one we often have heard is that people tried zTNA. They did a pilot. It didn't work out. It was too complex. And, again, they reverted back to what they had to their insecure, VPNs. So with that, I will hand it back to James to talk about how you can get unstuck. Yeah. So, you know, as Marcel mentioned, there are a bunch of different reasons why these projects may not work out. And so as organizations have started down the SaaS journey, we at CloudFlare have gained more experience helping those customers get back on track with these types of projects. So, you know, typically, we see customers structure their modernization in three overall phases and maybe 10 steps in general. They generally have two goals. So the first, that reflected in that top row, is to simplify infrastructure and tooling that they have accumulated in the past, and they wanna consolidate those capabilities onto one composable Internet native platform. And the second goal, kind of reflected in that bottom row, is to extend visibility and controls and to adopt that those types of zero trust security best practices between users, resources, across environments, on any device, in any location. So Topgolf, round simplifying and consolidating, bottom row, kind of maturing their, security posture. So we're gonna go deeper into this progression in a minute, but the reality is you can tackle these phases in in steps in any order. And in fact, if you've already onboarded a service like Zscaler Internet access or ZIA, maybe you've actually started by focusing on phase two. Maybe ZIA is protecting your outbound Internet traffic because you wanna defend remote users. Makes a lot of sense. But we often speak with CIA customers who have been hesitant to deploy Zscaler private access or ZPA for a lot of the reasons that, Marcel mentioned. And that is really what we're, you know, referring to in that phase one when with modernizing access to internal apps. And as a result, those organizations are maybe still are struggling with legacy remote access tools like their VPN. So, again, Marcel, I'm curious based on, you know, your experiences, what are some reasons that we've seen organizations maybe hesitate with their, ZPA deployment? Yes. When we when we talk to customers that have tried to roll out ZPA, I think the number one reasons that we hear is is the higher complexity of ZPA. Right? In contrast to ZIA, where you roll out an agent and then all the processing is done on in the cloud, rolling out ZPA requires you to, connect the users to your private application, which is in your data center. And that means you have to spin up virtual machines called the app connectors. And so the overall complexity is higher and people just experience the same ease of deployment that they are known for from ZIA, which is unfortunately not the case. Also because, ZPA was built after ZIA, and the user interfaces and APIs are completely different. Right? Zscaler tried to, fix this with the Zscaler experience center. So if you have deployed that, you will see that it looks like one single UI. But if you look deeper, you will see that the policies of ZIA and ZPA are still completely different, and they also follow a different kind of methodology. So, this is an increased complexity that we've heard so many times. We also have heard about incompatibility with legacy applications or VoIP connectivity are very common reasons. In general, Zscaler's on ramps are more limited. We offer a lot of on and off ramps and, including, you know, connecting, your, infrastructure to us with fiber cable, very advanced things that you can do with us that Zscaler does not offer. We also have heard about poor user experience. You have to pick up your ZTNA traffic as soon as possible and send it through a private backbone so it makes its way to the private application as soon as possible. Zscaler's answer to any performance or latency issues is often to deploy additional hardware on-site. And then you're going back to the old days where you had to manage appliances, which in return will increase complexity again. We also have heard that the commercial inflexibility is an issue. If you want to make ZPA fit your requirements, you often have to buy a lot of costly add ons, and that will balloon the cost up and maybe create a a budgetary issue down the road. And that is that is another, frequent complaint that we have heard. Yeah. Thanks, Marcel. So, you know, Marcel highlighted a a number of different reasons why organizations may struggle to roll out ZPA or may be hesitant to kind of tackle this first phase in their journey. So if you are maybe one of those types of customers, maybe you have CIA, but you've been hesitant to kind of focus on these remote access use cases, we often work with customers on tackling this first phase. And this has really been our a a focal point, obviously, since the pandemic as organizations have adapted to distributed and hybrid work. So let's kinda walk through what exactly we mean with these various, steps within this first phase. Now a common starting point is clientless access for a priority group of users and applications. We're gonna go deeper on this shortly, but for many ZIA customers, this is the best way to test out Cloudflare's approach in the short term without disrupting any of your existing deployment. And over time, though, we do see a lot of organizations look to deploy a device client Cloudflare's device client so they can scale their this zero trust approach to more users and apps. And this is here where you can layer on checks for multifactor authentication, device posture, other contextual and risk signals. And for many, as we mentioned, if they've kind of struggled to, adopt the zero trust approach, for many, the goal is to replace those types of legacy, VPN appliances such that they can enforce default deny identity based access across all their apps and environments, including non web and legacy private networks. One of the kind of core findings we've encountered in helping organizations make progress on this journey is the value of starting clientless. So, actually, last year, we commissioned a survey of some 200 decision makers in security and IT. And one of the key findings from that study is the strong agreement that doing things in agentless way and a clientless way can really accelerate zero trust adoption. It's a good fit for a lot of apps and users that are high priority to protect. It's certainly simpler to deploy because you don't need to juggle another device client, and it's cert and it's easier to scale. Now prioritizing these types of agentless use cases can help generate quick wins and build your comfort with a vendor, with the zero trust policy building approach, and kind of help you figure out, you know, your comfort level before introducing potential headaches with managing another, device client. So we encourage clientless use cases, wherever possible. We want them to be used more widely, and we actually often see it underutilized as a stepping stone. In fact, you know, I think we think that you can start, modernizing remote access right away for priority apps and users with CloudFlare without installing software on devices. Now if you do that with CloudFlare, we're gonna be able to run-in parallel with CIA to enforce identity based zero trust rules across your web apps, browser based SSH, RDP, and DNC environments. In fact, many of Cloudflare's largest deployments started with this cloudless approach before scaling up. So let's look at some, you know, examples. First is securing access for contractors and third party. These are folks often on unmanaged devices, often have a a really popular starter use case. And this was an early step for Indeed, the job website before they decide to roll out to their full workforce of 13,000 plus users and removing their VPN entirely in just three months. Another popular use case is securing privileged access for developers. These are folks who are interacting with sensitive infrastructure often. This was the starting point for Canva, the Australian design company. And actually building comfort with Cloudflare's approach help inform their decision not to build their own identity and access management system in house, and that they could just work with a provider like Cloudflare. Another popular use case is simplifying IT integration during mergers and acquisition activities. So this makes it easier to onboard newly acquired teams and get them started and onto your systems without doing a full network merge. In the case of Delivery Hero, this is an organization who, you know, expands into new territories by acquiring brands. This helped them scale significantly, after facing significant demand, post the pandemic, you know, with with delivering, food. And last but not least, we often see organizations experiment with kind of new cloud based clientless CTNA approaches for apps as they are migrating those apps from on prem to the cloud, and that makes more sense oftentimes than trying to retrofit an on prem VPN. So for example, with the investment company EQT, clientless CTNA helped them move away from their on prem Microsoft Active Directory, avoided having to retrofit their own custom proxies and so on. Now all these companies have ZTNA deployments with Cloudware in the thousands now, and addressing these types of use cases early on with a clientless approach help them build momentum as they scale further. So I'm gonna turn it over to Chris now who's gonna show us how these clientless capabilities work. And in particular, we're gonna look at a demo. He's gonna talk through how you might run CloudFlare in parallel with Zscaler. Thanks, James. Hey, everybody. So I'm gonna take us through what these clientless deployments look like and how they interoperate alongside Zscaler. And one important thing to remember is that despite their motions to unify their user interface, Zscaler ZIA and ZPA remain separate products. And it it's for this reason that, Cloudflare's ZTNA can very, you know, fit very easily, alongside a ZIA deployment. So as you can see at the bottom, ZIA is still serving as a forward proxy for your users. So it's handling Internet traffic, enforcing security policies. But at the top, you can see Cloudflare access or Cloudflare's DTNA protecting and securing access to your internal resources. So that means your web apps, your, you know, browser based non web applications like SSH, VNC, and RDP, even while the Zscaler client is running on your machine. And this matters because it's important to reduce the amount of complexity in your security environment, which means, you know, avoiding things like setting up VPNs for your contractors, third parties, developers, and users on unmanaged devices. And Cloudflare access can even isolate applications that is protecting and enforce DLP policies to the browser in a way that does not interfere with any web protections that the Zscaler client may be applying at the time of policy enforcement. So this clientless option works well because Cloudflare only needed you know, it only needs to manage your organization's DNS, in order to protect your application. So when someone tries to access your protected domain, their traffic is on ramp to Cloudflare's Edge and we apply any and all security protections that you've set for your application. Let's take a look at what that looks like in action, over on the next slide. Alright. So here, we have an origin server that's running a Wiki application on Docker. It's listening on port three Thousand. Currently, it's a private server, so there's no public access. There's no public IP. And right now, as as you can see, we're creating a connection from the server to CloudFlare. We're creating a secure tunnel between the two of them so we can access the server. We're gonna go ahead and name it WikiServer here. And we're using Cloudflare d, our standard app connector. So with this, we're gonna go ahead and, after we get the confirmation that the connector has installed, I do wanna point out that we provide these pre baked commands over here that are you simply need to throw them inside your command line, and then we're going to instantiate the CloudFlare tunnel and then connect it back to CloudFlare. So this process is very easy, when compared to the setup process, for, you know, a ZPA connector, which is going to require a lot more, you know, a lot more in console and in platform work. Basically, just to even, you know, get it connected, let alone proxy the application across. So right now, you can see all we have to do is define the Wiki, for our, you know, our subdomain for our our our domain and then define the service that's gonna be proxied across this FQDN. And, again, with the, you know, with the general process of doing this with ZPA, we had to run the exact same process. There would be many more steps involved, and you would also, like, very likely need a more powerful device depending on how you're using this, whether as a connection broker, or, you know, in general, you know, as as a Here in IDPs, you can see that we're supporting a large range of IDPs in our environment. We already have a number of them configured, and we also support, like, any OpenID Connect or SAML compliant IDPs as well. Now this is important because it's a bit more difficult to set up all of these IDPs, inside of an instance of ZPA. So, whereas Cloudflare is going to allow you to apply as many IDPs as you want and apply them a la carte to all of these applications. A big problem, you know, that you'll run into, with a, you know, ZPA deployment is going to be that you it's going to be a lot harder to do so. Mainly because, first, you need to contact support in order to get, multi IDP, you know, in order to enable your account for multi IDP. And in addition, you are not going to have be able to apply as many IDPs as you would with CloudFlare. So over here, you can see, in this application, we've applied it to Okta. We've also applied it to, a one time PIN, over email, which again is an option that does not necessarily exist over, over Zscaler. We're gonna go ahead and add an application definition and, down here, we're, you know, telling it to apply to this, apply to this policy here. So here, we can look we're we're actually leveraging and creating rules here, to be able to allow people to get into the application. So these are reusable objects, application. So these are reusable objects, much like you would see inside Zscaler, and we're leveraging object groups. We're also, you know, we're leveraging identity groups, and we're also saying that anybody who's associated with this policy is going to have to have multiple device posture rules applied on top of that. Now coming back to Zscaler, you don't actually, define that device posture profile, inside of ZPA. You would actually apply it inside of the Zscaler client console, and that would need to then be applied and referenced back inside of ZPA. Again, just adding another, you know, another con you know, another management screen, and another point of failure, that you would then need to make sure is up to date whenever you're creating this application. But now we're coming back here to the sign in page. I'm signing in just as an employee of the company. I'm I'm using multi factor authentication as I set inside of the policy. And we're gonna be able to get access, protected access to the Wiki. Alright. So now we're gonna go back and we're gonna define a policy for contractors in the same application that we just created. We have a lot of ways that we can identify contractors. We could say, for example, have an email list where we're managing a list of approved emails for contractors. And this is a list here that would be maintained directly inside of CloudFlare. So I'm gonna go ahead and add that to the policy. We can also leverage, you know, group membership like before, you know, if you have if you have an existing group that's used for contractors, or you could allow authenticated users where their email address domain matches a specific domain that's associated with firms that we're hiring contractors from, like KPMG, or PwC. And, again, one thing I do wanna point out here, is that contractor access with Zscaler does get a bit more complicated, mainly because you do need to manage, it would need to have access or be able to sign in via the IDP, the one time PIN over email. Again, that that, you know, that contractor access becomes a bit more difficult, to apply. And over here, you can see it's just as as easy as adding an IDP option. So any user who authenticates and and matches any of the selectors, they're gonna be able to get access to the Wiki. But we're also going to apply another security policy on top of this to isolate this application, meaning that's going to be rendered and served, from an isolated browser that's spun up at Cloudless Edge. So for this policy, which is, you know, targeting people who are outside or external to the organization, we want give them access to this Wiki, but we also wanna be able to control what all happens or what all they can do once they're inside. Again, securing access, based on the context of the request. So we have, you know, somebody signing into the Wiki and we're also gonna notice, that once they sign in and, you know, once they get the code, they're gonna have the data protection controls that are applied on top of their session. So what that means is we can control certain activities like copy paste, uploads downloads, you know, printing or even keyboard interactions altogether once the user, has access or what, you know, access to the isolated version. So we're gonna go ahead and, you know, we're gonna try and print. And we can see that website printing has been disabled. And that's just, you know, the nature of it being rendered inside of an isolated browser. Also, the keyboard itself is entirely blocked. Alright. So at this point, you've seen, like, what our client list ZTNA solution looks like in action. And another reason we wanted to show this was to point out that CloudFlare does have a superior ZTNA solution to Zscaler ZPA, thanks thanks to ease of use with installation, you know, stronger identity and device posture integrations, in general, the performant nature of its global network. So with that, we've reached the end of our demo section. I'm gonna pass it off to Marcel, and he can continue with modernizing access to CloudFlare. Great. Thank you very much, Chris, for the demo. So now I will talk a little bit about the, benefits that Cloudflare brings to the table right when we talk about modernizing, access. As you have seen with this demo, Chris was able to connect a private application, to the employees within how long was that? Maybe two or three minutes. So, we offer very simple management with browser based access for self hosted apps and infrastructure. We have one intuitive API and one intuitive UI. Right? There's no need to deploy the identity and sign up for the experience center in one API. We are already there. We also don't need to deploy any virtual machines. As you have seen, this can be a simple command on the command line. And we also offer a proven faster and and better performance in first and third party testing compared to Zscaler. We run-in over 600 data centers across 330 different locations worldwide, and every service is delivered from every data center. Zscaler is not in the business of running networks. And and one place you see that is that that we are the number one most connected network in the world, while Zscaler is not even in the top 25. So we also offer one unified network. There's no CloudFlare one, CloudFlare two, or CloudFlare three, right, where you you need to find your tenant. There's only one network. And we also have a composable cloud native architecture. What that means is and I can give you an example. We run one of the largest web application firewalls in the world. And when the need came up to also protect private applications with the web, we could just reuse what we had built over the past ten years, also for private applications. Because we're composable, this means all our services can be shifted around, and they can be rearranged to actually build new services. So in this case, a web for private applications. So, today, you can leverage our battle hardened web application firewall that you know and love from protecting your websites. You can use the same for your critical internal applications. And we are also the only vendor that can connect and protect all your traffic. Right? Zscaler often talks about attack surface reduction. We can actually eliminate your attack surface. Right? That means we can, isolate your email servers, your websites, your internal external apps, APIs, contractor portals. Everything has to go through CloudFlare before it's allowed to touch your resources. So this is very, very interesting because it allows you to apply least privilege access, scan for threads, block data exfiltration at all these different points. And we are one of the few vendors that can do that across the board, not only on the SaaS and zero trust side. And last but not least, we are also, we are not forcing customers into cookie cutter contracts. Right? We are we want to be collaborative. We want to be a partner of choice and also help you, achieve your you know, be successful on your SaaS journey. And so we allow you to have flexible contracts with us. We don't, enforce certain up, ramp times, and we don't want you to have licenses sitting on a shelf. We actually want you to be successful and have a working, ZTNA implementation. With that, I will hand the word back to James. Thanks, Marcel. So we really focused on modernizing remote access types of use cases. We focused a lot on clientless use cases. But, of course, that's often just a starting point for more mature for maturing your approach to a zero trust. So we wanna share one framework to consider where you are and where you wanna go. So here we have two rows that have show a good, better, best maturity levels for both user authentication and resource authentication. I think we can all agree that the maturity levels at the far left in, you know, are bad. We don't wanna rely on usernames and passwords. Users should not have access to everything. And we think every organization should at least try and make progress and get to stage two. Right? Access should be based on identity, and at the very least, you should get groups of authenticated users to apps based on certain IP ranges. And, of course, the aspiration is to keep moving right. So for users, I've been enforcing MFA, device posture, those are best practices, and you can add more contextual risk signals as you go along. And for resources, this can vary a bit. Maybe sometimes you you'll just be comfortable with authentication to private network IP ranges, or maybe you wanna go kind of all the way in in terms of doing per resource authorization based on individual host names. Now, ultimately, we think that how far you go on this continuum will vary on a number of factors, but we want you to have these types of considerations in mind. And in working with you, we would, you know, guide you on some specific steps to help you move between each stage. Okay. So we focused a lot on remote access already. The next phase is really focused on defending against cyber threats, protecting data, replacing complex costly point solutions that have traditionally been in the workspace security space, so your buywalls, your off prem proxies, and so on. And maybe if you are a ZIA customer, this has been your focus to date. We're what we'll we'll kind of talk through what this approach has, look like with Cloudflare. So in step four, we see a lot of organizations look to strengthen their web and email security. That means temp deploying technologies like a secure web gateway, remote browser isolation, email security. Maybe that means starting with DNS filtering to protect outbound Internet access and then enforcing more granular controls with HTTP and browser isolation rules. In step five, organizations want to prioritize mitigating the risks of data exposure and leaks. So in particular, this you know, they wanna regain control over different sprawling SaaS environments. So they do this with technologies like our CASB and DLP. And over time, you, in step six, kind of extend visibility controls across all channels. So web, email, SaaS, cloud, private app environments, you have those types of SSC controls. Now if you have DIA or maybe some other web inspection going on, chances are you started down this path. Right? And if that still makes sense for you, you know, godspeed. But we also encounter a lot of organizations who have some of the frustrations that, you know, Chris and Marcel have highlighted, before. And so they are looking for opportunities to migrate to CloudFlare and make progress. So we're gonna talk through a couple of examples of of where we've encountered that. Starting with THG, they're a global ecommerce retailer based in The UK, founded in 02/2004. They currently have a a large workforce of over 7,000 hybrid workers. And they've begun a VPN replacement project with Zscaler, but were struggling. And in particular, they were frustrated by having to switch between two separate interfaces for ZPA and CIA to configure their policies. Now THD had already been using Cloudflare to support some of their ecommerce projects, and so they saw an opportunity to consolidate more of their tech stack with Cloudflare by migrating in a way. And in this way, it could kind of unify how they were securing access to private applications as well as, their Internet under one dashboard. Now that migration process was accelerated by some automated tooling we've built that allows you to transfer ZIA configurations to Cloudflare's environment. We like to call this our VScaler toolkit. And so, for example, in just one week, we helped migrate several dozens of Internet filtering policies covering some 100 plus content categories and custom URL list. Now even if, you know, you as an organization don't go with the exact outputs of that descaler tool, this automation gives a sense for what zero trust could look like with Cloudflare. And over time, they've been able to extend those policies to, you know, their their hybrid workforce. Let's look at a second story. This is Applied Systems. They're a US insurance technology company. And when a new c suite took charge, they began to reevaluate their architecture. In particular, they want to move their away from their traditional Cisco AnyConnect VPN, and they were at the time using Zscaler ZIA for web filtering. Now that VPN was too slow and risky, common complaints, and developers were also frustrated that Zscaler was blocking some of their work critical, apps and destinations. And like THD, they wanted to unify access controls on one platform. So applied systems, again, similar to THG, had built familiarity with CloudFlare in other use cases, including protecting their public facing websites and applications. And so doing more with CloudFlare allowed them to replace eScaler and Cisco and and consolidate. It it it made sense as a strategic direction. And like other customers, Applied Systems also started with with clientless e t n a for their web apps before layering other layering on other controls. And to their credit, they've been able to automate most of their deployment, taking advantage of Cloudflare's API first approach, and in particular, automating how they onboard new users, new applications, build policies with Terraform. A third example kind of focuses on one of the top children's research hospitals in North America. This organization was an early adopter of another, SSC service with Netskope, where they were using them to secure access to internal apps, but they have grown frustrated over time to help. Maybe the quality responsiveness of their support was kind of in decline. They were also frustrated by some of the slow and unreliable connections. And they were also beginning a multi year strategic plan, which saw ambitions which had ambitions for the hospital to expand their footprint globally. And ultimately, they saw CloudFlare as a stronger, more agile partner for those longer term ambitions. And in particular, they value the breadth of our network relative to, in this case, Netskope to serve those international expansion, opportunities. So every few months, they would expand our zTNA for another handful of 500,000, 1,500 users at the same time as they were unwinding Netskope. We're gonna go a little bit deeper with one final example. In this case, this is a global hospitality leader, and they had begun modernizing remote access, but we're still struggling with managing Cisco any connect, their VPN, and Cisco umbrella for web filtering. And juggling these two separate tools is too complex, and leadership wasn't really convinced with the their cloud strategy. So the hotel had begun the ZTNA project with Zscaler for some third party contractors, but they encountered couple of different issues including a frustrating login experience for their end users, some frictions, when it came to relying on third third party cloud infrastructure. And so the goal was to integrate those Cisco point solutions and find an opportunity to consolidate with another, player. In this case, they had phased their project in two stages. The first was migrating off Cisco Umbrella, which was at that time due for renewal. Month one was of that engagement was focused on planning, so aligning our technical requirements, business priorities. We think that's a very critical step for large scale projects like this. Month two, they started deploying our device client and testing out how it would work to do, for example, DOH encryption or making sure our client worked alongside Cisco's. Next, the hotel began its pilot, so they chose a random sample of some 2,000 users. This was at 10% of their overall target population, and this approach helped them kind of gather feedback and uncover issues across teams, regions, and environments, such that by month three, they could begin a larger scale rollout to their 20,000 users for web filtering. And by month four, they were able to begin uninstalling, Cisco umbrella and kinda make sure Cloudflare logs were flowing to their analytics tool. The next phase was then focused on unwinding their VPN. So this meant doing it required doing a little bit more with our CloudFlare device client in this case. So fully proxying traffic, enabling device posture checks, connecting tunnels from apps to CloudFlare. In those next three and a half months, we're focused on different configuring zero trust policies in different test environments, enforcing things like identity, destination protocol. And similar to in their earlier phase, the pilot in this case took another random sample of some 2,000 users and 4,000 users. And in the and in this case, IT administrators were prioritized so that they could have fast track troubleshooting with a group that was very inclined to share feedback regularly. Now over the next three months, they were able to unwind the rest of all of their solutions, including the remaining contractors who were still on Zscaler such that over the course of fourteen months after they started, they could commission decommission their VPN and safely turn off Zscaler. Now our initial conversations with these customers started very small. For example, we had started by just focusing on maybe enabling some of their housekeeping staff in one region to access that apps on shared iPads. But as the hotel began to understand the breadth of our portfolio, the flexibility of our architecture, and how we worked with them, they felt confident to go on this longer term journey with us. So I'm gonna bring Chris back on who is going to help us kind of, show how some of these capabilities we've highlighted, work together. So I'm gonna turn it over to him, and he's gonna do a live demo for us. K. Thanks, Jay. Okay. So So here we should be able to see our secure gateway policies that we have here. We have basic policies such as, you know, blocking unacceptable content such as things that match to content categories or blocking high risk websites. But what we're really gonna focus on is in HTTP policies down here. You're gonna see, a, you know, a number of ways that our services interoperate, in order to provide a comprehensive level of protection for your outbound Internet traffic. So first thing I'm gonna, point us to is redirecting users to chat g p t, which is our top policy. If you have users that are trying to access, you know, unapproved or unsanctioned AI applications, in this instance, you know, anything that's not chat g p t, then they'll be redirected, to the AI of choice. So you can see this working here if I try to go to a separate AI tool, and you can see that we have the 301302 redirect, over here to our, you know, to chat g p t. In addition, I wanna point out that we also have, generative AI prompt controls in place. So if we have, you know, the application is in chat g p t, and the DLP profile is in, financial information or personally identifiable information, we're going to be able to stop this. So if I were to and since we can see this policy is in effect, if I wanted to do something like copy a bunch of customers, you know, credit card numbers or Social Security numbers in and ask it to parse them for me, what's gonna end up happening is we're gonna see a network error, and we also see a helpful message from our work client down here saying that we cannot send sensitive data to AI applications. Now this is a custom message that we have here mainly to point out that this is the policy that was matched to, but also to give users a little bit of coaching as to what they are and are not allowed to do. In addition, I'm gonna go back here and apply another control on top of this, which is isolating generative.ai applications. So, much like you saw, with the, with contractor access, to the Wiki application, we're going to be able to isolate any web page that users get to, in a headless Chromium browser that's spun up at Cloudflare's Edge and we're going to be able to apply data protection controls on top of that. Now a very interesting thing here is since this headless browser is spun up in Cloudflare's Edge, it's still gonna be subject to all of the gateway rules that you currently have active here. Meaning that your, you know, the DLP and also the, the redirect policy for for generative AI, those are all still going to apply. So I'm gonna go ahead and spin up an instance of JANGBT. Okay. Now we can see that chat g p t is spinning up inside of a headless browser as we can see with the isolation toolbar here. And I'm gonna go ahead and try and paste that information in again. So we can see the paste here is blocked. Now if I were to do something like, you know, tell me, you know, from this file, any Social Security numbers related to Chris, then based on this, we can see that a network error has also occurred, and also we get the same message that we cannot send sensitive data to AI applications. Now we do have copy paste and uploads and downloads blocked here, but the nature of it is if you have, you know, personally identifiable information or financial information that's going into the AI and it's also happening through an isolated browser, the DLP profiles are still gonna be able to take effect. And what's really important here is that it can take effect without the user having a CloudFlare client installed on their device. Now how exactly does that work? Well, in order to do that, I'm gonna go ahead and switch off the client here that we've been using, and I'm gonna turn on a Zscaler client. As you can see, we have the Zscaler client service enabled. And I'm gonna go to just a basic policy, like, if we have users who go to, like a gambling site. Right? So poker.com. And you can see here that, again, despite the fact that we have the Zscaler client enabled, we have a policy that has redirected this traffic over to our instance of Cloudflare. So this is a headless Chrome you know, this is basically a stand alone clientless, isolation tool that anything you search inside of it, you'll be able to, search inside of an isolated browser. So over here, if we just go into, you know, once I sign in Once I sign in, we're gonna be brought to that stand alone is you know, browser isolation tool. And, again, anything I search inside of here is going to show up, inside of, you know, inside of our policies. And, again, because I have the policy over here in DNS policies that says we're blocking an accessible content, it's also blocked here by CloudFlare. And, again, I have CloudFlare policies applying, but I don't have a CloudFlare client on my device. Showing how these two tools despite, you know, both being, you know, you know, both being SaaS, you know, SaaS solutions, they can both interoperate with each other, in a way that is nearly seamless. In addition, I think that it would probably be good for users to leverage Cloudflare's browser isolation mainly just due to the strength of its network, and, you know, due to the large propensity of cloud you you have the single pass architecture, so every Cloudflare data center is gonna be running every service, as opposed to Zscaler, which is only gonna have about, you know, a few dozen, hubs out of it, you know, all of their points of presence that are gonna be able to serve, isolated applications, which would, in effect, create, bottleneck or, you know, perform performance bottlenecks for your organization. Alright. So with that, that is the end of the demo section here. And we're gonna pass it back off to James who can keep it going. Thanks so much, Chris. And, you know, just to round this out on, we're gonna talk through our kind of final phase, which is really focused on, you know, taking all the on prem boxes, those routers, firewalls, and TLS lines, and kind of shift all that network connectivity to the cloud. The goal is to replace all that overhead across your branches, your offices, your data centers, and other locations. So step seven is kind of if you think about the experience of walking into a WeWork or a hotel, just being able to grab a spot, connect to Wi Fi, Companies want to deliver that same type of simple, consistent experience wherever their employees are and adopt this kind of lighter weight, more cloud native approach. It's commonly known as coffee shop networking. So here, the SaaS platform, not the on prem firewall, is providing the protection. And with step eight, organizations try to manage different form multi cloud infrastructure, whether that's in AWS, GCP, Azure, with greater consistency and flexibility. So for example, Cloudflare helps organizations automatically discover and manage native functionality of a bunch of cloud resources like virtual private cloud, subnets, VMs, and so on without requiring any additional compute VMs. Eventually, in step nine, the goal is to eliminate all your private network circuits, especially those kind of expensive MPLS lines, and instead, route all site to site traffic through the SaaSy platform. And, of course, you can always get more granular with your zero trust approach. Your networks and environments are gonna keep changing and expanding. So in this final step 10, the focus is just to continue refining those policies to be even more adaptive to the context and evolving risk factors. So I'm gonna turn it over to Cliff Chris one final time who's gonna talk you through how these different on ramps work at a technical level. Thanks, James. So here, you can see just the, you know, more holistic views of ways to on ramp traffic to CloudFlare. We've been highlighting clientless access for unmanaged devices, you know, to reach apps and for proxy and traffic. But at these later stages, again, those common things that you're gonna be able to see here is the, you know, device client on ramp. So that would be, just have you have an active device client that sends traffic over to CloudFlare. But if we're looking at clientless options, that's going to be things like your IPsec and GRE tunnels and network routers, that are basically going from your firewalls and connecting to our WAN. So here, the WAN connector is going to be, the most recommended approach, but our WAN is compatible with a host of third party services for more manual integrations with VPNs, with your firewalls and SD WANs from, you know, any of your hardware vendors like Cisco, Fortinet, Palo Alto, and more. But further, when you on ramp, you know, from your cloud infrastructure like AWS or GCP, Azure, you know, any of the highest, you know, hyperscalers, you can take advantage, of newer orchestration abilities. For example, you can automatically discover cloud resources like virtual private clouds, subnets, virtual machines, route tables, and routes. You can centralize how you manage those resources to Cloudflare's dashboard. And as a final option, you can connect your network infrastructure directly with Cloudflare, the, you know, CNI, Cloudflare Network Interconnect, rather than using the public Internet, for a more reliable and secure experience. And like you saw, inside of the demo, we do have the standalone browser isolation tool, and anything that search inside of there is being searched inside of, in this Chromium browser in your environment, which means that that traffic is automatically on ramps, to CloudFlare. So with that, you know, between those, between just our standard clientless access version where you have the FQDN that routes to a tunnel record. Those are going to be, you know, our clientless deployment options, to be able to flexibly secure access to applications to users in any context. Thanks, Chris. So just as we wrap up, you know, we've talked through three overall phases. We've kind of talked about how it may make sense to to start with clientless, to run CloudFlare alongside your web proxy if it if it happens to the ZIA, and use that as kind of a way to build familiarity with a new vendor before progressing with these other stages in the journey. But, of course, customers have started with any of these three phases, any of these different steps, and the exact order of operations is up to you. You know, with our breadth of our portfolio, I think we are in a unique position to help customers start modernizing their their security in a variety of different ways. So as we kind of wrap up our prepared remarks, I'm gonna bring Marcel back. You know, please, if you have any questions, add them to the chat. And in the meantime, you know, I think, I'm curious for, Marcel, your profile perspective. Our hope, obviously, is that organizations choose CloudFlare for as much of that journey as possible, but we recognize situations where it may make sense to take a best of breed approach or kind of a dual vendor approach. And so before we kind of wrap up, any final thoughts on where that best of breed approach may make the most sense? Yes. Thanks, James. There are probably many practitioners on this call who would, think that adding yet another vendor, right, to your security stack will increase the complexity. And it goes contrary to the no call for vendor consolidation, which has become very popular in cybersecurity today. My thoughts on this is that if we go back, right, CloudFlare started fifteen years ago to make the Internet better. And ever since, we have offered all our services to millions of, Internet users free of charge. Right? And this has forced us to build, products that are easy to use and easy to deploy, and intuitive because, otherwise, we'll be drowning in a lot of tickets from our free customers. So, together with that, we we also built a platform from the ground up. Right? We we built it in 2010 with, you know, not even knowing what services we would run-in 2025. Right? SASE being a great example. So all our service are API first, fully composable, and run-in every location. And so this is a strength that that, you know, CloudFlare brings to the table, and ZTNA is is one of the best ways you can actually experience the difference. It could all your employees in a fast and reliable way, without deploying hardware, virtual machines, or suffering from, you know, disjointed policies. So as a former practitioner, I support vendor consolidation, but it should not come at the expense of security. Right? Every day you continue to run a VPN is is a day too much. That makes sense, Marcel. So thanks so much for for, you know, know, my my colleagues for joining and sharing their, perspectives. Thank you for joining this webinar. If you wanna go deeper, we really recommend our SaaSy reference architecture. It's linked here. This includes a lot of technical steps and best practices that can help you figure out where it makes sense to kind of onboard or can maybe potentially migrate to CloudFlare as you progress through those SaaS journeys. So that's a good resource, or you can get in touch with our team. There are a bunch of different ways you can get started, including architecture workshops, different proof of concept trials, and business value assessments to help you make the case internally for your project. So thanks so much for your time, and I hope to see you on another webinar. Bye.