Cloud Native Computing Foundation KubeCon + CloudNativeCon China 2018 (Shanghai), 21 Apr 2021

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Securing the Perimeter - CFCR/CFAR Chain of Custody With CI/CD Pipelines - Keith Strini, Pivo

Description

Join us for Kubernetes Forums Seoul, Sydney, Bengaluru and Delhi - learn more at kubecon.io

Don't miss KubeCon + CloudNativeCon 2020 events in Amsterdam March 30 - April 2, Shanghai July 28-30 and Boston November 17-20! Learn more at kubecon.io. The conference features presentations from developers and end users of Kubernetes, Prometheus, Envoy, and all of the other CNCF-hosted projects

Securing the Perimeter - CFCR/CFAR Chain of Custody With CI/CD Pipelines - Keith Strini, Pivotal

Continuous integration (CI) systems automate the building of working code and Continuous Delivery (CD) is the movement of that working code into the hands of end users. These two concepts enable companies to move with effective velocity into new markets. However, because of the pervasive nature of building, testing, and delivering into production, even with immutable containers and a kubernetes secure by default posture, CI/CD pipelines could be used to severely compromise the entire software delivery process. Chain of Custody need to be established and enforced. In this talk we will cover several aspects of ensuring chain of custody, including transmission security, rotating developer keys, signed git commits, independently reproducible build verification, signed release artifacts and run time authority for immutable containers achieving secure an end-to-end chain into production.

To learn more click here: https://sched.co/FuJb

A

Hello, everyone- this is sergeant ding, and this is Keith and today we're very happy to present. This is securing the perimeter with the CFC R + CF, AR, +, +, CI, CD pipelines, so CFC R is a container, is Cloud Foundry container runtime. They say this is a foundry continued runtime is a distribution of kubernetes and CFAR is a cloud foundry application runtime, which is based on like open source distribution of cloud foundry, and you know Keith and I.

A

We are both from pivot, oh and we work for products called people, cloud foundry and we provides commercial products based on this. Cfc are open, source and CFA are open source and before before, were a process to the next slide.

A

I want to introduce ourselves, I explained what wait, what do we do in pivotal and what we are so my name is sharon ding and based at Dallas, I'm, a pivotal foundry principal solutions, architect and I've been working with a pivotal for five plus years before that I worked for vm work, so it's always infrastructure as a service platform as service. It's been a long time and I would like Keith to introduce himself a little bit. I told him how to introduce himself.

B

C

This is gonna, be really bad. I want to try this right so.

D

Washer washer g zu g, chi-su Liza, you are Johnny G, yeah.

A

That's not bad right. Okay,.

D

A

So Keith Lee is also a Solutions Architect pivotal. We are on the same team. We are basically like focusing on North America region and we basically provide operational assistance and guidance for a pivotal. Our foundry we deploy cloud foundry to you, know the products on the customer side, it's more like post sales and beside that we also build balanced customer team. We want the customer operational team to trade, their developers as a customers.

A

We want them to treat the platform itself as a product and we teach them a lot like agile practice and how to build pipelines with a you know, both infrastructure platform and we.

B

A

Our customers, with a application deployments because.

A

Services, you feel the pipelines you want to continuous delivery. Candido gather feedback from your customers. That's basically like what we we're doing here and you further comments. Yes,.

C

So the big idea that we want to talk to you guys about today is basically we want developers to do what they do best, which is write application code.

C

But what we don't want to do is sacrifice any of the security aspects, and so what we're going to show you today is how we walk through from code to commit all the way to deploying a production quality application and how we offload these things, with continuous delivery pipelines, automation within a different phases of that delivery process, and bring that all back together and into how you integrate this. With your security policy, yeah.

A

So next yeah, we won't talk a little bit of shape the mindset and basically like we want assume the continual threat of compromise traditionally like when Enterprise dealing with a security attack right, it's more like static like, for example, they secure the parameters with a firewall and injects a place, some antivirus software into the system like for us. We want, like borrow some concept of disaster recovery. We want to tree the platform that the security attack as a you know like the it's kind of like disaster right. So what happens when you have a disaster?

A

You want like a recovery to the previous point of the you know, either by using recovery point objective. It's basically like the time that you wish to recover to the otherwise RPO is how much time you take to recover that in pivotal, Cloud Foundry, we got a product called parsh. We we try to convince our customers to tree the platform to be more ephemeral and we're basically like a lot like even banking system.

A

We recommend them to repay the whole environment right tree to be ants ephemeral and as long as they have the high availability which our products provides. I weighed this car, the VN like every 24 hours and doing some rolling repaving when we call repaving we'll be paving all the VMS within 24 hours, so assume like you've, got like a security attack.

A

As someone like even inside the employee, they play something like that: a newer virus all like they got like some meat mine in the middle of an attack to call your certificate, but because you just discard the whole VM right, you've, trees, everything to be ephemeral and its repaving everything, and that's actually like just the the minimize, the security risk and the security attack, but that's the concept of repaving. What we want like keep reiterating in this talk, maybe you want to chamber a little bit more here sure.

C

So the big idea here is that the hardest target to hit is a moving target right. So the idea is, if we look at a recovery point- and we set that, for let's say every 24 hours, so we won't lose data any longer than 24 hours in the past and we're it takes us. You know, say four hours to roll off the entire environment, then we're going to continuously do that, knowing that we have a safety net there between our objective and our time recovery.

C

And so then what happens is a cyber attack of some sort, whether it's an advanced persistent threat or a virus is just another disaster and we treated just like disaster recovery, and so this allows us to continually move that environment target, to make it very very difficult for somebody to actually attack it through whatever different means that they're going to do that.

A

Yeah, so we definitely want to talk a little bit how we can secure the pipeline, but before that I want what is the standard delivery pipeline? It's ready for application. So pivot-hole like we like to automate everything. We regard our products about conkers and we using conkers to support pipeline right. So TV CS is specific to the version system. Kill hobby is a standard, TV CS and you can have like the CBS and other. You know version system.

A

So, basically, like you checking the code to your DVCS and and then you start, Union tasking start building your boundaries and your containers, docker images and then push to artifactory or like darker registry, and then assume you have a platform as-a-service here and you basically like pool the images or boundary and deploy to different environments. We call this is build ones and promotes two different environments. This is cool right, so you can continuous delivery as long as you check in the code but like because we got this automation.

A

Sometimes it's like, if you have, like one point of you- know: virus, for example, malicious code, someone checking to the github, because you have continuous theory, and it can be, you know up into your production environment. So how do secure your products? Environment is very critical here.

A

So we're going to explain like some like prototype or principles, to secure your pipeline and make sure you don't get like something, some malicious code into your production, so the first one is defining the supply-chain thread and no matter whether using github or CVS, something that something about that they always have. This sign the commit feature right basically like a developer. They have authorized keys right. This is a kind of like all authentication and authorization and github definitely have this feature.

A

You should use that and you can rotate and we would like the developer to rotate the keys right so basically like this is a standard, cryptography and developers son the commits when they push by using their key and their private key and the pill system who like when you got along holding.

B

A

The from the github you busy doing the signature verification, because public with public kids were pre-installed over there and the same concept about another data is your boundaries of darker images. You want like a sign the commits by using the builds you during the build phase and the artifactory you when you're pulling the data from the artifactory or darker image you you basically like make sure the signature verification. Also over there and again you don't have to do that by ourselves right.

A

So, if you're ever using torque or IO, and you don't to use the darker content and trans, that's you expose yourself to something like security risk, I, always trying to. If you're working on a before Enterprise, using darker, counting trust and I. Remember we have some. You know: harbor harbor insert order registry. You know it's kind of like enterprise level, docker registry.

A

It has like a notary and whenever you push the docker registry with a darker content, trance enabled and the developer will Sun the image with their private key and then, when you pull that it has to be a Sun. If it is not Sun, then the polling will be fair. It will failing right, so this is basically like defining the supply chain thread by using signature, and this is a very standard, cryptography way of the secure pipeline and your data anything so.

C

The primary reason why this is labeled phase one is because this is the first place of which you can start doing rotation of those keys to make sure that the development force, that's contributing to that codebase is active. So, like you know, if you have several companies that you outsource your development to and you as a company as though as the provider that's gonna, take those and put it in front of the end-user unit.

C

Don't necessarily know who's been hired, who's been fired in those different companies right so and then, if they fire somebody, but then their key or or you're not using keys at all, then they can and they may still have access to contribute to your code, repo, which then means that they can inject that malicious code. So phase one is being able to rotate those developer credentials so that they have to check in with you on whatever that policy is, let's say you rotated every 30 days. That means every 30 days.

C

You know exactly who is still hired by that particular company and who's authorized code contributors and you can trace every line of code back into the sign commit. So you know exactly who wrote the code if there is some sort of malicious intent, so this is phase one where you can actually impact your security on on the very initial part of that supply chain.

A

Yeah, the other one is the limiting C ICD breakout Explorer. There are two points: I want to talk about. First one, is you wanna like secure data in transit right, so we talked about like the signature on the static data, but like for the security, the data in transit, you want to enable TLS and like sometimes, we also want to like using mutti I asked for the mutual trust before between the client points and several points, so I always use TLS.

A

Some people argue that you know this is the internal system, but you also want to avoid the man-in-the-middle attack and also there may be some inside the employee that can thicken do security attack on your system, so always using TLS and mildew tiaras and the other one is the the local environment, the credentials to be local environment right so limited your credential to local environments, for example. You don't want, like the the build stage of deploy to tasks to have the credential overlap with the deploy to QA. I saw a lot available.

A

They love to check in, like you know, it applies credentials sometimes, and also they are using property files rather than you know environmental variables so always trying to limit your credentials to your local environment. That's a good practice. Anything yeah.

C

So why we call this phase. 2 is because this is the second area where you can have the big impact on your actual security. The big impact is that you have no transference so, even though shoutian or I might be able to use stuff in production and do stuff and test and maybe leave when we do code contributions. What typically happens is, it seems, like common sense, like only have the credentials for that specific environment. But what would happen is like. Oh, hey, it's key. Let's go ahead.

C

He should know he has permission to do that stuff from production. Let's give him permissions on his account, so I sign in once and suddenly I can access every environment. Well, if I get grumpy, that means I can affect everything all the way into production and that's a bad practice. So the very first thing is like no transference between those properties in the environment. That's a one key area that you can definitely lock down and keep that regulated.

C

So that means, even if I can talk somebody in to give me a credential on the build team or talk somebody into giving me a credential attest. I still can't get to production. You see so that I am very limited in what I can actually do and then the second thing is using TLS and transit. You're gonna rotate those those certificates out right.

C

So those things will always be reactive, but you also can IP whitelist between those particular environments, so that you only have point-to-point actual authorization to actually get in that environment through another specific environment, and so this also limits your breakout potential and this particular thing. So this is why we call it phase 2 of impact on your security.

A

Yeah, the other thing is, if application or data integrity again, this is building into github or torpor registry and basically like the github. When you push the code, if they will conceal the hash a pound like what kind of like files you have by the way you pull the data, it's it's going to compare the hash again and then do some verification. This will guarantee you that you know if someone plays some malicious code and the integrity will be there. Along with like a signature verification, we can have a strong application integrity. Yes,.

C

Oh yeah, this is this is really good, because this is what we're this is the heart of what we want to talk about today, which is the application code as the unit of currency that moves through the entire system. This is what we're going to both secure it up to this point, but also, as we bolt on other things, make it a consistent experience across that applications, evolution and what that means is that my coder doesn't have to know anything about security downstream.

C

They don't have to know anything about what it looks like in production and what they need to do. They don't need to know about. You know how to configure middleware services or databases. All that stuff is we're about to show you how that stuff is handed to them. Well, what we're doing is we're now taking delivery of the code that they've committed into our particular environment and going to move it and graduate it up into production, and so looking at that unit of currency over the next slide.

C

That's what we want to talk about is the steps, so these are very generic steps right. These are everything that happens to be able to put that level of maturity around an application code. So we start with code, that's built, tested in security scans. We already.

C

We already know that the code is safe at this point in time, but we need a lot of other things to it that actually make it viable, and so some of those things are first, you know we're talking about cloud scale, so we need to know what are the available host so workload capacity? How do we balance this across all of our other tenants on that particular piece or within our company, all the other application workloads that are on there? So we look for available hosts.

C

We configure on top of a harden container, so what the hardened container we'll talk about it in a little bit, but really what it is, is it's a stripped-down operating system that is only there that has just a necessity necessary libraries in order to run the container itself. So this means that it's actual attack a threat surface is very, very small as well compared to a and operating system.

C

We might only have 35 to 40 system libraries in there that could be exposed and they're very easy to monitor and fix an update with CDs and patching, and we want to grab all the backing services. So, like I said, we don't want them to worry about how to configure the database how to cluster the database. Do these type of things. We just need to know that they need a database to use so we're gonna, inject the environment, variables and the connection logic into the actual application automatically for them.

C

We're gonna, pull that application source code. We're gonna, create the package, we're going to configure the dependent service and then we're gonna schedule that add that container to start and then what we got to do is start looking at the production quality stuff. So we need to know like if we want to run in a high availability configuration well, we need to consider things as a platform.

C

We need to consider things like what, if an entire availability, it goes down what if an entire region goes down, so we need to scale up a number of instances across different availability zones and different regions so that we can survive an entire region going down, so we're gonna handle that for them automatically we're gonna look at setting the route, the SSL termination, the configurational load balancer all of this stuff. If you think about it, this is applied against every single application that goes out there.

C

So we have a consistent like we don't forget to do a route or we don't forget to set the firewall against it. We don't forget to do the SSL termination. This is all done automated for us, and the developer doesn't know anything about that, so they don't have to concern themselves with. How does SSL termination work? How does the load balancer work? We take all that stuff away from them and make this all the way automated until we get the app into production.

C

Now these same steps are applied, whether it's the cloud foundry application run time or the kubernetes portion of this Cooper now is we're still catching up to some of this automation stuff, but we hope by the end of next year we you should be able to accomplish all these 16 steps in either in either particular offering whether it's application run timer or container runtime services I just.

A

Wanna say like we already achieved this with a CFA, our physical, our foundry distribution. We are working a lot on premise: distribution of a peak es which is kubernetes very wise. If you ever worked with it at the vm work, we we partnered with the VMware on SXT basis, like it's, provided a lot container to container container networking network policy and also distributed firewall, and we are trying to like and like more and more features in the in the automation way to to create a secure pipeline to make sure we deliver.

A

You know the security features on the platform as service of on fly.

C

So, as you look, we we basically packaged everything we needed to get the application of production so now we're in the run sign a portion of this. So this is phase 4. This is how do we impact the run sign and just quick shout out, and that of Seth is a platform architect for us. I borrowed these this diagram from him, because it was very useful to talk about some of these things.

C

What, if you look at this, so this is, you know a notional of a container running there and inside this container we have things like user name, space, mountain name space, and what these do is allow us to restrict user scopes so that, even if you compromise one of the users within the container you're very limited in what you can actually do, that user only has a very few permissions that it can actually execute against.

C

We also do filesystem isolation in there so that you can't impact other aspects of the filesystem and then, of course, it runs on top of the hardened container image, which is what we were talking about before. So we have both the stem cell, which is the VM level operating system. That's been stripped down to just run containers, and then we have the container root filesystem and there we do would like an overlay on top of that, which means that the application itself, the binaries, are in a very small read/write layer. They have very little impact.

C

If you break out, you really can't go anywhere, there's there's nothing that it can actually do because of the way we've restricted where it runs. We've provided everything it's a run, and so therefore it itself cannot actually go and do anything outside of that. That particular piece like.

A

It's just one: toggle is beyond the stem cell right. So the way PCR works with a stem cell is so we retrieve the stem cell as a VM templates right. We have a securities hardening pipeline, it's running every day. It's if there's a CVE comes in and this pipeline we will just apply the patch and it's january, the stem cell VM templates for each infrastructure as service.

A

It applies GCP or vSphere, and this kind of, like stem cell actually applies to both kubernetes and Cloud Foundry distribution, and they are the only thing is like we for Cloud Foundry. We have this root filesystem, but since, since you can have this, what I call the the docker, and so you can pull the docker image, is now like as hard and as Cloud Foundry.

A

You have to use the just one way of root filesystem, that's why you have to make sure you secure the target registry in the future, and we have this Enterprise Drive or hard. It's very important as a commercial offering affricate.

C

Of them, the next slide so continue yeah continuing on. We look at network namespace, and this is how we do network isolation with port virtualization. So the reason I'm showing you all these different pieces of this is because these are all the things that, as so, the Cloud Foundry application. Runtime is a little bit more mature than the container runtime, but we're working to make the same level of parity inside of kubernetes so that you have the same level: security within the actual container itself and then ruas containers.

C

Obviously you know, anytime that you have a bridge between your root vm and your container vm and in any of those accounts you have a problem because you can escalate privileges. You can break into the actual root, vm + f8 those privileges and do some more malicious stuff, so with Ruis containers. What this means is that there's no concept of mapping between a user in the container and a user on the vm, which means there's no way to break into the vm and elevate those privileges.

C

So these are also concepts that we have to think about as we move into the kubernetes world and how we're going to productize these types of things for the same level of enterprise level, maturity and then a part armor and set comp, restrict those untrusted processes and in system call filtering, so they even if you spawn, even even if you got out somehow and you want it- to spawn a process to do something nasty to the to the environment, you're going to be blocked by the actual Apple arm or piece which will prevent you from doing untrusted processes and they container itself.

C

So this is really important. All these concepts are applicable, whether it's the CFAR, the CFC are, and so this is this is how making sure in the runtime that we're still maintaining that level of security once our users start interacting with these applications.

C

And then we talked about this a little bit, which is the continuous zero down si de patching because of these contracts. We're really trying to get to is this particular model. This allows us to swap out vulnerabilities if we discover a vulnerability like a virus itself or some sort of malicious piece, we're able to rotate out that stem cell without impacting any of the actual user base right. So as we do this canary style deep deploy, we can actually move the user access to a different, a different VM, a rotate out.

C

The stem cell wrote that out the runtime rotate out the root filesystem and then all these are patched moved workloads back over to the left side and then do the same thing. And then we just keep doing that until we actually get across the entire enterprise.

C

And so again, your user experience doesn't go down, but then you're also still focus very focused on the security aspects of of that stem cell and of the root filesystem itself, and so this same concept is exactly what we're doing in on the PKS I, which is the kubernetes offering which is looking at those root file systems. And how are we going to rotate that out from from the actual doctor-doctor registry?

C

The next part of this is the other. Continue studies advanced persistent threat. So what an advanced persistent threat is it's the type of code that gets into your system doesn't act right away. It sits there for a while, and it learns about your system and, as it gathers more and more information about your system, then it gets ready to do an attack. It says I've now discovered the IPS I've, now discovered the types of file systems or operating systems, I now to discover all these different things and now I'm going to launch an attack.

C

So how we get rid of this is, we assume a zero trust network right, and this is what you want to look at- is all your components that make up your distributed operating system, whether which it, whether it's a platform as a service, whether it's your own distributed, offering what you're doing is through mutual TLS between the components we ensure that you know you're stuck on one side of the other. We rotate out those credentials.

C

We can repave every system component in there and then so, even if that persistent thread is in there, it's got a very small time frame in which it can learn what how to attack your system and then suddenly we wipe it off the map and then we've created a brand new instance of that, and so you have to start your attack all back over again to get back in there to do something bad and so, like Chen, said earlier.

C

If we're setting this for a 24-hour period, that means you have less than a day to get in learn about the system launch an attack to some other system component. Even if you did you're gonna be stopped by the TLS command, a TLS aspect of that, and even if you did, that, 24 and 24 hours, you're gonna, be wiped off and have to start again, and so this is phase five.

C

This is when you're bringing everything together: you're repaving your entire stuff, you're rotating, all your credentials out, you're, updating all your your certificates, so that they're new and current you're you're getting rid of any disgruntled users that happen to still be in the system, because you've rotated those things out and you've brought it all together into like a final vision of how to continuously move this target for attackers to try to hit and here's the zero duds model rotating. All those system credentials jump.

A

One slide ahead: yeah.

C

I just kind of combine a ball and then, finally, if you look at what this looks like so for us, this is what we do. So you know wash is how we actually manage all these things in a consistent fashion, so whether it's Bosch or at some other release management. That's you know totally up to you guys. Well, what it does is, if you can see you know you combine in a sixty with like the Software Defined Networking piece that does a container level security perimeter security.

C

Then inside each of these big boxes, you did the component level credential rotation and the repaving of those platforms, and then you can and then you can continuously update the the vulnerabilities through the actual virus piece, and so all these things give your platform operations team, a single user experience of how to actually manage this thing from a security aspect. It gives you consistent security model across any of the type of workloads that you're dealing with, including the the networking piece as well. So it's pretty much.

C

You know you've thought from the beginning, from your developers all the way through to production and how you're going to manage that in in terms of the runtime you go to the last slide and so to bring this all together phase one. You know the thing you can start with his aggressive rotation of those developer credentials or key developer keys phase: two is rotation of the environment, credentials, endpoint, IPS and dynamic management of those in points.

C

So this is your in your TLS in transit, IP, whitelisting from point one to point two and then rotating that search, so those are always consistent are always current and then phase. Three is the continuous verification, the application integrity. So this means that, even if somebody attacked me in the middle of that process like post, build but pre deployment in operations, I would immediately see the hash, inconsistencies and I would know the application was tampered with and then once I get into runtime.

C

Is that making sure that it's only something that's authorized to run in there and if it ever broke out restricting its movement? So they can't go anywhere else, and so that's phase four and then phase five is bringing that all together and doing all the different pieces, including me at least privileged on the container and renewal, those authorizing credentials. So all these brought together.

C

You really think, through from your initial developer, commit all the way into your post application and in deployment and then how you're going to continue and manage those workloads and I think that's it. You guys have any questions.

C

Quiet, that's not good, sir.

B

C

Developer signing signing of commits yeah, it's very straightforward, actually github has that feature. Doesn't it yeah, I think.

A

You're talking about the image is signing right, so there's a there's, a framework called top right. It's a tu F is updated framework which is, is it's are being adopted in the darker world, so like dr. Doyle or the other one is called harbor I think you know other vendors. They have docker registry. Provide that feature as long as you enable darker content once or if you Google darker content. Trust I probably cannot Google here despite the darker content, trust and it you know, once you enable that you started' pull push.

A

The image is automatically signing the image and using this update framework, the other, the other terminologies called notary service, notary services, kind of like implementation of tough and yeah. To answer a question directly like this is our ability into talker, though, all like you know, commercial Darkrai registry as Harper is very transparent to developers.

C

Yeah your biggest level of effort and that is managing actual keys themselves, so you're gonna have to create some sort of internal company process that does it on a on a cycle like hey every 30 days. We regenerate these keys and then we hand these things out. How do we make sure that these things are done or, if you're doing it within the image signing it's? How are we going to actually manage those?

C

What kind of cadence are you going to manage those keys on, but the actual how it works is very, very straightforward, because it's built into a lot of the technologies today, it's just more the difficulties getting your company to enforce that policy, so that it is actually done on a regular basis. Any other questions.

B

B

A

B

Maybe you want to repeat.

A

The question sure.

C

So so what he was asking is, so all these things are nesting kabaddi from a whole listing.

B

C

Which things could you implement first as low-hanging fruit, and some of these things are so some of them are dependent on the type of company that you are right.

C

So, whether or not you rotate your developer keys is dependent if your developers are in-house, it's maybe not as necessary, right because you'll know when you hire or fire somebody if they leave or just become discontent whatever you still want to protect against it, because the disgruntled employee, you may never know right itself, post attack, but in general that's not what happens right so, if you're a company that has everything in sorts, maybe you want to focus on something else now, rotation of credentials or certificates can be very difficult.

C

It depends on the type of technology that you're using right. So you know it's hard for us to sit up here and not be biased right. So if this worked for total right, so cloud foundry has a stuff built in what other technologies do as well.

C

So if your technology already has it, then I would make sure that you're taking advantage of a technology that already has that, if it does not, then things like you know how you're managing your containers and runtime is probably the most critical right, because this is your if you think about it from a user perspective, where does your user give you personal information through your application? Where does your? Where is the most exposure to the outside world through the runtime in the application?

C

Where is it that most likely folks are going to attack your production environment? So you really want to start with your production environment and making that secure and then work your way back into the less threat bottles. You always look it up from the threat where's the greatest threat. The greatest threat is wherever it touches the internet and the most people have access to it right. That is most likely where your threat is going to come in each of the edge cases which are like insider threat.

C

Those are those tend to be edge cases that don't happen as frequently, so you always want to approach it and like it, the escalating sort of thing most most important, but most likely to lease likely and so I would really focus on your actual runtime. It would be the first thing that I would try to try to work out how your end time works.

C

Yep, no problem, any.

B

Other questions.

C

Come on don't be shy, I mean.

B

C

Mandarin, it was terrible, so what nobody's gonna speak I was gonna, do worse than I did other questions.

B

D

A

It was out so you he was asking the demo. You know for this security I. Do oh no.

C

We don't there's no like we have. We do this all the time with our customers, but we don't. We don't have a triple yeah.

A

Because these a 35-minute bridge.

C

Is a timewise we don't haven't had like right now, we're at it yeah.

C

Yeah we have lots of repos that this is built in that you can go check out for sure. If you go on.

B

C

Can get it to you? Just email us like on the presentation. You can see our emails just email somewhere since your repo that you can see all the stuff in action.

D

B

B

B

B

C

So both so what this is is this is born out of Sochi I have been doing this for a long time with a lot of fortune 100 companies, and so this is borne out of our experience when working with those. So some of these things like, if you look at phase four like yes Cassandra already, does that, but this isn't a cloud foundry Talk. This is a. This is the type of thing that you have to do some way.

C

If you choose the technology like say Papandreou, it's already done that's great, but if you want to do yourself, that's fine, but you need to think about this. This kind of thing, but you know we come in, and we do a lot of these things with our customers as an additional value that we provides them because of our experience with all these companies understand so yeah. This isn't, like a must add. This is more like experiential, like what we've done in the past.

C

That has worked out really well like there are security teams, a look at this and they say ah that makes sense. This is what we want. You know- and this is how we do it so that that's what this is born out of, just to kind of give you guys ideas of the things that you have to think about from a Brett perspective, for free to your.

B

D

Question: okay, that's it thank.

C

You thank you for uh for coming.