Zscaler Brings a New Approach to Experience Management

welcome to episode one of z cast for 2022 everybody i hope everyone had a great holiday break and it seems we're right back at it now uh before i get going though i want to give a quick shout out to my media partner e-week as always z-cats are done in conjunction with the e-week e-speak series uh today uh to kick off this year i'm delighted to be joined by dewal sharma vp of product management for the red hot cloud security company zscaler uh welcome to wall and why don't you just give us a little intro on yourself thank you zeus and happy new year uh hi everyone this is double and vp of product management of z scalar being with company for almost 10 years i focus on our one of the key uh growth initiative rescaler digital experience and few other emerging technology product areas and happy to be here today yeah so this is an interesting topic because as i mentioned uh you know zscaler we know you as a cloud security company but uh we're gonna be talking digital experience and uh back in november uh of 21 we actually talked about zdx a little bit and when you announce some enhancements to the service uh so uh given your strength in cloud security uh why did zklr decide she's gonna decide to build a monitoring product like this and how does it fit into the overall z-scatter portfolio yeah so that's that's a good question a lot of our customers ask us but if you look at our growth over the years last decade uh z-scale as a platform we have focused on uh making sure our customers get better security no matter where their employees are whether they sit in office or they are outside the office uh we act as a security exchange sitting between users going out to any destination whether they are internet or private applications a lot of our customers adopt z-scaler as part of their broader digital transformation journey and when we think digital transformation uh one of the key intent is to embrace cloud mobility and uh if you are adopting these technologies and you are going for better security it is also very very important to have good user experience we focused on the user experience aspect for many years by building point of presence across the globe with 150 data centers today but uh as we as we started thinking more and more uh we our goal was to make sure we provide the shortest path to these destinations for our customers but also uh there are a lot of issues that we started observing that were happening not due to let's say our cloud presence or uh not due to just isp issues on how people connect to us but due to local routing problems uh isp first mile connectivity problems or sometimes just due to local users device issues so by virtue of how we have built our platform we have an agent which just tears traffic to the cloud sits on the user device we sit between the users and the destinations they are going with the cloud we have built uh with massive presence across the globe and with our private access service we actually have a a uh presence in front of customers applications so we sit in the entire data path and we thought we are sitting in the right place to build a end-to-end user experience monitoring platform that leverages our point of presence and is very complimentary to our security offerings so it is very very complimentary to what we are we have been doing in the security space for a long time yeah now monitoring is an interesting area because it's very broad very deep lots of incumbent vendors and the typical enterprise has typically deployed three five ten maybe even a dozen or more monitoring tools already so um what gaps does your product fill and why would a company consider zdx yeah so that's that's the question we asked ourselves when we started building zbx so motivation for zdx just to give you some uh background backdrop on it about three years ago when we started thinking of building zdx first thing i did first i looked at all the support tickets that have come into z scaler uh from our customers so about 20 of all the tickets that come in our sport queue come in the area of traffic forwarding or how people connect to us and 80 of the issues that we observe were due to the fact that uh not a z-scale problem these were like either last mile connectivity problem or local issues of the customer's environment or the device issues now first thing we started looking as is we ourself have been using off-the-shelf monitoring tool in our cloud operations and what we always found was they were very reactive in their approach to your point availability and performance monitoring market is has been around for more than three decades and there are established vendors in this space but as we have seen in the world of security there were tons and tons of point monitoring or point security tools that have been getting consolidated with uh secure service edge or zero trust network access based approaches we thought there's an opportunity to bring this consolidation in the world of monitoring there are very silo tools to your point uh a typical enterprise is using dozens of tools uh like there is network performance monitoring tool which network teams use there's end user or laptop desktop monitoring tools which is used by workspace teams there is application monitoring or ap em tools and all of these tools are living in silos and what we have also seen is as you adopt more and more sas services every sas sweater will stick you with their own monitoring tool like microsoft will give you a skype monitoring tool and then you have trust portals from different vendors so it became very very difficult for organizations to see when a problem happened what is the impact and most of the time they were reacting to it so we saw an opportunity to converge this space uh by consolidating a lot of point monitoring tools into a broader digital experience monitoring platform so a lot of traditional monitoring capabilities kind of are collapsing into what we are building but uh in a way we fit into the broader monitoring ecosystem well um now we analysts you know we love to create acronyms and things like that and good folks at gartner have been recently using a term called sse which stands for security service edge uh and what they're trying to do with that is describe the evolution of security to a platform does zdx fit into that yeah so uh sse is uh very interesting from our positioning perspective to be honest uh for a very long time as you would know z-scaler was positioned in secure web gateway but uh as the leader in this space but what gsketter has done over many years is uh uh cons on the security side we said there's no need for individual firewall there's no need for a standalone sandbox or let's say for a standalone browser isolation service right so we are consolidating these and as you think about what we have done is and again we introduced the concept of cloud-based security running on the edge right not running deep inside a data center and building pops to just steer traffic to that data center so uh it conceptually z-scaler has been doing what is called secure service edge for many many years from day one uh to be honest now uh so we see a very good alignment with that positioning uh and as i mentioned right for us uh the digital experience monitoring or zdx product is basically uh helping customers realize the full potential of that security because if you get the best security in the world and your user experience is suffering as a result of it your end users will not be happy right so a lot of our customers who are adopting zdx are seeing immediate return on value in terms of uh not only getting better security but if there is a bottleneck somewhere because remember a lot of customers who are transitioning to cloud security come from a very legacy architecture so their way how they send traffic to us or how their enterprise networks have been built over many many years there could be bottlenecks that they are not even aware of it right so our intent with zdx has been to make sure that those bottlenecks could be identified and customers know where they need to act on so it kind of becomes a very cohesive uh framework where you are providing better security but also enabling better user experience by telling them where things are and that's very important to build a comprehensive uh security edge architecture that g-scaler has played for many years yeah that uh juxtaposition of good security and good performance has always been there and it seems like there's always been a trade-off but it sounds like you're trying to address that now i was uh looking through your last uh earnings announcement and i believe the commentary on zdx is that it's z-skater's fastest growing service now so i'm curious to uh understand what's led to that growth and how you're selling zdx with the broader z-scale platform yeah so there if you think about gdx uh as a uh product line right so it's an independent uh product line the way we have built it right so uh it's a like we have this color internet access we have the scale of private access zdx is affiliate product line but it's built on the same platform which is very very important so we're not selling siloed point products we made we made a lot of architectural decisions on day one that it would be integrated or an integral part of our customers journey so if you think about uh one of the key thing is uh every enterprise hits putting more agents on the user's device right as z-scaler five years ago when we built our agent we made a decision to have only one z-scaler agent for all our products and services so first thing we did is all the instrumentation from a monitoring perspective that we had to do we built in the same agent so as a result of it some of our very very large customers with 70 000 users 100 000 users who are deployed with our internet access service they need to deploy another agent or virtual machines or browser plugins they basically got the license within few hours they had like all these users experiencing being monitored and this was a big big win from their perspective now another thing which we have done is the way we uh have uh launched the product into the market is you can uh of course customers can buy it on their own but from a go to market perspective we focus on having this as part of our broader platform so customers who are buying more security services can have zdx bundle with it those customers who are already on security services they can buy it as an add-on and those customers who are starting on the security side because one thing which we'll talk a little bit later is also the different kind of personas because typically we sell to uh office of cso cios and networking teams get involved but zdx goes beyond that right so uh at times our uh champions who are buying our security services don't look at the zdx because the time crunch then uh we also are giving them a standard version of zda exceeded in with our core services of z-scaling index and privacy access so they can get let's call it a taste maker for zdx and then they can deploy it and then this immediate value and then they move to the next version of zdx yeah so can you drill down that a little bit i know there's two versions available you have zdx standard and zdx advanced uh what do i get with standard and then when would i upgrade to advanced yeah so zdx standard uh is uh basically uh bundled with our z-score internet access and z-scale private access edition so any customer who buying these services uh beyond the business and editions will get a zdx as well the standard version one of the key thing or decision we made is we instead of giving you 10 seats of zdx uh with 100 seats of z z scalar or giving a subset of those we actually give zd extender for every single user that for whom we are buying our security services and the immediate value of that is the way you look at the way you think about gdx it is basically monitoring three variables that could impact end user experience network endpoint application right with zdx standard we are still monitoring every single user's device health right like how their cpu is doing what is that how their memory is doing like is their device healthy is this connecting the right wi-fi we allow them to monitor up to three critical applications or set up six probes like network level probe application probes whether these are internal application whether these are sas application it doesn't matter so they can set up six probes with this and they can uh see data up to last 48 hours so this helps them proactively troubleshoot things but also get a good understanding of the values edx is delivering for for the entire user base we have many many customers who actually have deployed gdx standard for their entire workforce they're seeing emulate value and some of our uh recent large gdx advancements have come from customers who actually were deposited gdx standard but they were ready to move to zdx advanced where we unlocked the full potential of zdx where essentially what we do is we allow them to monitor uh uh tens of applications we can monitor up to a set of 200 probes which essentially is 50 applications you can monitor then you we have a integration uh with third-party tools and uh we have uh api we have uh web hooks to integrate with other ticketing system we have alerting framework that gets unlocked also we recently introduced ucas monitoring so you can basically do api based monitoring of your microsoft teams and zoom tenant as part of it and there's a lot of additional things coming down the future in gdx advanced that will keep making it more and more valuable uh just to add one more point to it so one thing which we also saw in an opportunity we saw based on our customer engagement is that a lot of customers have in the last two years have adopted uh office 365 or microsoft 365 services and they had especially customers who were moving from or traditional office to microsoft 365 are struggling in terms of understanding where the bottlenecks are with performance so we also introduced zdx for m365 package which includes key office application like outlook teams and sharepoint pre-packaged so customers can just deploy that for m365 and then they can move to zdx advanced as well yeah yeah now you would uh that's interesting because uh um you initially said that a lot of your sales are run through the cso organization but what you were just talking about actually are use cases for it operations and help desk teams so correct them wrong but it seems like those are also organizations um that will benefit from zdx correct yes absolutely so uh again uh it's all part of as we build the product we will uh we are learning ourselves as well we engage with customers and we learn where the zdx is being adopted some of the mistakes that we made earlier i would say zdx still uh being around for just over a year now as a generally available product but in first few months when we made zdx available to customers our security champions loved it because as you can imagine every time a user is having a performance issue who gets blamed security right so security used as a tool to prove their innocence but i think what we realized is the full potential of zvx was being realized for organizations who are giving it in hands of network operations then we're operationalizing it with service desk and workspace teams so then we started seeing an opportunity to show the value of zdx to these potential uh teams within the organization and they loved it for that reason because as i mentioned most of the time monitoring tools are very reactive something bad happens you get hundreds of tickets there's a cost associated with the ticket and then uh you look at monitoring tools and they basically act on it we kind of flipped that on the head the way zdx works is it instead of like you setting up thresholds and understanding what is perform indicator of good versus bad performance we basically uh baseline the user experience no matter where they're coming from for a application they're using and then we look for deviation if there's a deviation in the zdx uh we use a scoring principle called zdx score my score acts like a thermometer if you are showing symptoms we can look at what potentially is going wrong and then depending on persona uh whether you are from network team application team and pointing you have which self telemetry that you can look and your typical level one level two people can uh look at gdx data can understand what is the problem where is the problem and then they can understand why that problem is happening so the mean time to detection comes down significantly a number of tickets that get opened when a large incident happens uh at that time you can send out advisories to your employees much sooner so the ticket volume doesn't come in because every ticket has a cost associated with it so we have seen some of our these uh uh tremendous benefits from our customers who have adopted gdx in uh last six to nine months yeah well it seems like some of the benefit there is a move away from that concept the mean time to innocence and actually trying to fix the problem instead of just proving it's not your own fault right so yeah you know that that's something it's very good at it seems um now in november uh you added ucas monitoring and if there's one market that's been as hot as cloud security it's cloud collaboration right so how does your ucas monitoring work yeah so uh again driven by customers a lot of demand on this just to be clear right uh our customers were able to monitor application like zoom and teens uh even before we introduced the ucas monitor just to double click on that like uh understanding the network path to let's say team's front door or to zoom's data center was there in zdx from day one we could monitor zoom process on the device and tell you when zoom process is consuming a lot of memory for example those things were there in the product what we have introduced now is getting more specific meeting level details like who were that attendees on a given meeting uh how was individual attendees uh uh meeting experience and another thing which we did is uh as as you can imagine like these collaboration applications are very different from typical web applications right so the kind of metrics that you really want to monitor are also different for these applications for example in uh for web application we compute the score based on a key metric like page fetch time which is like how long it takes to load the page in front of you uh when you are accessing that application but that metric is irrelevant for ucas application so we basically uh enhanced our security our scoring logic to have the score for ucas application computed on moscow moscow is much more relevant and it includes factors like like audio jitter video latency sharing latencies and what we do in this case is we basically have we give customers an option to connect to the call quality data from microsoft teams and zoom through apis and we basically enrich the network path and the device monitoring we are doing with the uh meeting level details so let's say if i'm uh having a uh like on a given day if i do five zoom meetings we you will see all the meetings i was part of how was my experience on each meeting and who were all different attendees in that meeting and you can then look at moscow what's the audio latency high and the best thing is like uh you can then drill down and see the high audio latency was that result of a bad wi-fi connection or was that a result of a bad network connectivity from my home or was that just because uh there was an issue on the transport relay side of these applications itself so uh getting that single pane of glass all these stacks from a single user perspective or a geo perspective or a company perspective or in a single place and this simplifies the life of someone who is troubleshooting these issues level one level two people or even when this gets escalated uh very very uh uh com makes it very easy to troubleshoot these scenarios yeah you know so in addition to ucas you also wanna announce the monitoring of zero trust architectures and uh so how the heck do you monitor zero trust and uh why is that important yeah so uh that that actually has been a i would say that was the harder problem to be honest because if we go to the like the underlying essence of zero trust architecture is that you make your applications go dark right uh you re you those applications are and this is the problem by the way which a lot of our customers who are using traditional monitoring tools when security moved to zero trust architecture their monetary tools become completely ineffective right and so basically as applications are not visible from outside you have to do instrumentation inside and even then another challenge you face with these applications is these are running in your data center let's say server is running there if you don't put the right monitoring architecture you might just create a denial of service on your own servers by just sending tens and hundreds of monitoring probes on there right so that we had to spend quite a bit of engineering time to solve this problem uh and this was another capability we made generally available for our customers uh who have killer private access service so we basically instead of just sending probes from the agent we basically do a lot of smarts on the our cloud uh broker service which is essentially our service search where all the policies are executed and all the sdp gateways which we call as app connectors we kind of uh use the principles of rate limiting and aggregating the probes and then we also have some telemetry which is connected in real time as application connections are set up to these uh uh from the gateways to the uh applications and we bring this telemetry out in a way that we you still get that visibility from every single user to the destination perspective but without creating any architectural burden so we had to do a lot of work on our side to support it but this is one of the key differentiating capability that we have for our customers and uh to be honest something which i have personally seen and we have seen it in our own z-score environment recently also is that the private application data paths are much more complicated and complex than the internet bound data paths because you have years and years of legacy networking in place you do not know how your internal traffic is being routed and every single user traffic takes different path so if you are not monitoring these we have customers who have been finding lot of bad network decisions they make a decade ago or two decades ago which are still in place which have been causing latency and user got used to them but with zdx now they are tweaking these architectures to make faster access or give faster access to private applications as well you gave me a lot of uh you know interesting uh you know theoretical examples of of the ucas and zero trust and other types of monitoring do you have any actual customer examples that you could share with us though of of people using zdx yeah so there are we have a lot of them because uh every time a uh customer finds something interesting with zvx they actually send a screenshot of us hey look what we found right so uh i'll give you a few of them uh what let's talk about a specific macro scenario macro scenario is when there's a big sas application like office 365 or salesforce which you have an outage right so uh when salesforce had a big outage on their infrastructure back in may of 2021 we actually saw that a lot of customers were getting impacted as a result of it so and that was cascading over a period of time so it impacted the uh people in uh international markets first u.s people started seeing slowness in connecting their connections were timing out but for people outside us they were simply not able to connect to salesforce even our own employees who work in sales use salesforce living right so typically on a given day when this kind of incident happened uh even in z scalar our itt would expect about 200 to 300 tickets coming their way right uh but when we were deployed with gdx what we had we actually started seeing this deterioration in score within uh and salesforce basically after 90 minutes acknowledged that there was a problem with dns and it is impacting all the users there was a feed update from their cto on twitter and they also updated their health dashboard so we basically had identified this problem within first 20 minutes we sent out a field advisory notice to our our own employees saying there's a problem going on with salesforce we received about 10 to 12 tickets only where we typically would have received 200 tickets so the number of tickets that typically would have opened would gone down significantly and uh uh one of the way we have operationalized zdx is we have alerting framework the output of alert for like bad score for uh gdx across the globe for salesforce were sent to an uh operations team uh and they basically uh had integrated it with pager duty the output of that alert and they immediately got the escalation they created advisory and sent it out so the mean time to detection uh came down significantly as a result of it so this is like a macro issue uh we have customers who have done seen this with office 365 outages or recent aws outages and they have used zdx for those purposes the second scenario is more localized problem uh which is uh i'll give you example of one of our key executive for example he we were on a call on a zoom call and his uh zoom connection was dropping every few minutes like we see jitter and then it goes away and he's like zoom is having problems since morning on my setup right so we immediately started looking in zdx what's going on what we figured out is actually there is no problem on uh in local network in on the zoom application or in uh on the isp side he like all of us he has like hundreds of uh chrome browser tabs open there was some javascript process which was spiking the memory and the moment that was happening every five minutes zoom was running out of memory crashing and restarting that one so it was a local end user induced issue but someone who's not very technical would always think that a zoom problem right so those kind of issues we find day in day out and these are the kind of problems which are very hard to understand for it operations people as well if they don't get this single pane of glass uh where all the network application and point telemetry is coming in a single place um and one more anecdote i will add to it is we actually had a svp of id in a very large healthcare company who basically when he joined the organization ceo basically said uh one of the key thing we need to make sure is our employees are having good uh experience as they are working remotely and he rolled out a nps based uh or net promoter score based uh scoring framework to gauge employee experience they also adopted gdx and they started resolving tickets much much faster with zdx as a result of it within three months their nps score increased by 18 after adopting zdx to troubleshoot end user issues so these are some examples and uh we have a lot more we have written blogs on those things on our corporate website but it's always good to hear when some customer shares these anecdotes with you as uh zdx is helping them in their daily life yeah well the real example is a good one because it does show the value of having that end in view right so you know an apm application apm monitoring tool might see that where the npm one wouldn't and you know i think that's why troubleshooting application problems has been so hard historically um you you mentioned microsoft and zoom and uh i i know you also have a partnership in the with servicenow in this area how important are those partnerships can you talk a little bit about what you're doing with them other than slapping the logo when you're you know as a partner so so uh just to be honest right we have we refrain from doing slapping a logo kind of a partnership those are not very effective for customers we always take uh the if the the all the integrations we have done in the past in our security and uh in gdx we found that what we saw the kind of partner ecosystem we need to build is different from what we have typically done in security so two areas where our customers really ask us to focus on what is in collaboration area as you can see with ucas we are integrating with the provider apis and what we are also learning is people like microsoft and zoom are also very interested in the data we are seeing because uh guess what if you ask microsoft what is their biggest problem uh today uh for office 365 it's not security it's performance actually right so they they have done a lot of work in this area as well so but uh the kind of data that we are collecting is also golden for them so there's a lot of collaboration that's going on in this front with these vendors same with zoom as well uh now uh with uh another key area i would say eight out of ten customers that we spoke to have some itsm tool and servicenow is a leader in this space so we and we are we both are mutual customers and partners of each other and we've been working very closely with good exact alignment and product alignment with each other so one area that we identified was uh automating a lot of itsm workflows with uh an id1 workflows with servicenow i'll give you a couple of key use cases that our customers really care for one of them is in gdx we support this concept of live troubleshooting we call it deep tracing so when let's say when a analyst is uh analyzing the issue in gdx for a given user and they want to understand the current behavior and what is happening and they want to troubleshoot things there's a simply a button in zdx we call as deep tracing you click it it basically stuns a live troubleshooting on user's machine remotely for next 30 minutes or 60 minutes and we start collecting more telemetry on the user device what we are working on doing is uh automation is uh with the apis that we have and what uh itsm has is basically when a user opens a ticket uh we can automate creation of this live troubleshooting session without a use a analyst going and creating that uh troubleshooting session and enriching the ticket with more context second thing we are doing is uh which is more advanced capability that we are working on and we'll be running out pretty soon is let's say if an employee has opened ticket in last two weeks right and those tickets are open what we can do is we can look at gdx in the last two weeks and say in last two weeks tell me when this employee had opened tickets collect their zdx data and see when if they had a bad score at that time and automatically enrich those tickets with that context uh plus there is a lot of uh additional work we are thinking in terms so these are two use cases we are productizing now but there's a lot of additional work we are doing in terms of like employee self-help uh initiatives for example like a servicenow has a portal where you open a ticket we can add a widget from zdx to say hey you're about to open a ticket because every ticket has a cost right so uh before a user opens a ticket we can give them some some commentary on things that we observed that they can fix on their own even before they open a ticket so these are like some of the initiatives we both have identified we have started working on and this will help our customers immensely yeah i think more and more to the company's success is going to be based on its ability to partner actually so i'm glad to see you guys are doing that um and uh last question for you here uh wall is uh what what's the overall vision for zdx so what can customers of yours expect to see uh from this product uh you know this year and years to come yeah so uh without getting into uh roadmap what i would say when i think about the vision right so uh zdx the way you should think about it is right uh zdx as i mentioned is a pillar product line failure product line means that as a company we are making significant investments in this area uh we have expanded the team we are uh looking at uh expanding the where zdx is today we always start any product line we start without with a focus set of use cases the key use case areas with gdx have been around end user experience monitoring but uh the areas that we will take gdx would be beyond uh just focusing on end-user workflows for example uh monitoring from z-score data centers it would have been very easy for us to do those things uh from on day one but we focused on the harder problem which was end user experience so taking zdx uh the concept of cdx beyond end user experiencing on to the network paths and infrastructure and uh data centers uh in addition to that uh focusing on automation one one uh key area where we are actually putting a lot of effort right now in zdx is to uh bring more intelligence everyone talks about aims these days but one of the key advantage we have is we are collecting about eight billion telemetry points with gdx even today it is deployed by two million users uh to uh i would say more than two million two million status uh at least a couple of months old but uh the key point here is like we are collecting we're sitting on a rich drug of data what we have started doing is we have started uh applying machine learning models to this data which will help us understand why problems are happening so we can then uh build a lot of ai ops kind of capabilities into the product to say what is the problem where is the problem and with the click of a button you get kind of a small mini rca which tells you why that problem is happening so it brings the uh the time today it takes for a human analyst to understand those things will be reduced significantly as a result of it so intelligence and uh bringing bringing a lot of machine learning into it is something which we have started working on because uh we think there is a lot of applicability of ml in on the data we are collecting with zdx yeah clearly you guys are collecting just volumes and volumes of data and we almost need machine learning to be able to interpret what that means so i'm glad to see you bring that out uh anyways we're uh we're we're out of time here i think i'm gonna wrap it up there thanks for uh you know give us a little bit of insight um i you know i'm i always like products to begin with the letter z so i keep up the good work on zdx and thanks for that deep dive so on behalf of dewalsh uh sharma from z scanner i'm z is caraval from zk research uh thanks for watching and don't forget to click to subscribe you

This transcript was generated automatically from the video's captions and may contain errors.