Webinar: Enrich Your Live Streams With Advanced Analytics

Justin Miller:

Hey, good morning everybody and welcome to this Wowza webinar. My name is Justin Miller. I’m the video and webinar producer here at Wowza and today we are going to be talking about how to enrich your live streams with advanced analytics. So just so you know what the agenda is first, I’m going to introduce our speakers. We’re going to take a little poll just to see how people are feeling about analytics. We’ll talk about getting access to data analytics now in 2020 and some of the real problems with non-real-time analytics. That being said, we’re going to talk about how you can build a proper foundation to get real time analytics and the future of analytics and just almost sneak peek of our own Data Insights. After all, of that we are going to do a demo giving you an idea of what those Data Insights are like and we’ll have a Q&A.

Now, you can ask questions at any time. In fact, you should see that there is a Q&A area to ask questions. There’s also a chat window. That chat window is available for all of you to chat back and forth with each other, but the Q&A and a window is the window you’ll be using in order to send us questions. And again, those questions will be brought up and discussed at the end of this presentation. Just to make sure everybody can hear my voice and see my screen. Go ahead and in the Q&A window, just send a shout out, say hello just to make sure it’s all working.

So hello Crystal, how’s it going? Hope things are going well. Noel from Boston. Hey, how’s it going? Okay, glad to see everything’s working here. Now before I introduce our featured speakers here, I just wanted to take a quick poll to see how you guys are feeling. So my poll questions are here. Number one, are you currently streaming live video? We always want to know if you are or are not. And two, do you currently use Wowza to do it?

So if you could just answer these poll questions, I’d definitely appreciate it. Question number three, what industry are you in? And then question number four is multiple choice. We’re trying to get a little more information on what types of real-time analytics are you interested in. So for example, real-time data for streaming health monitoring, real-time data for analytics, real-time data through API, real-time data through dashboards and other visualizations. Doesn’t look like it sent the whole question there for realizations and the last one was real-time notifications based on thresholds for defined situations. So I’m just going to leave that up for another second. And while I am, I want to introduce Jamie Sherry. Jamie’s here, he’s the Wowza Media Systems senior product manager. Jamie, say hello.

Jamie Sherry:

Hello everyone. Good? Doing well?

Justin Miller:

And we also have Jim Hall from Fastly. He is the senior sales engineer there. Jim, great to have you here.

Jim Hall:

Thanks very much for having me.

Justin Miller:

All right, well I am going to end the poll right now and I don’t think we need to necessarily share those answers. I’m going to close that up and we’re going to jump right into our first slide. So Jamie, let’s talk about data analytics for 2020.

Jamie Sherry:

So, hello everyone. As you may indicate or know by being in this webinar, in 2020, data analytics have become pretty much mainstream, getting there, closer to mainstream if not already. It depends on the areas that we’re talking about. But certainly data has become more and more important overall in decision-making in business operations and in strategy.

And especially when we look at proving out a return on investment revenue, KPIs, these kinds of inputs become more critical than ever. As we know, quality and availability are very high today with streaming video. All the services out there we see with 4K and live events and always on and so forth and over the top TV are all examples of where you have higher quality, high availability. With live streaming, some of the difference there is you have this one chance to get it right. In a sense, you can do a live event and if you have issues, it can be challenging to recover quickly from issues. And in some cases your event could be over by the time you resolve those issues.

And if you have problems of a severe enough situation, you may not get a chance to do it again. One notable example, Not to pick on Hulu, but they’ve been doing the live debates over the years, this year and last year I should say. And last fall they had some challenges with one of them and they actually got a second chance, but that’s a rare case. A lot of opportunity and a lot of demand to do it right and provide that high quality, like what cable expectations are today. Right? TV for the most part just works. And people have those expectations that have continued to just work.

What we’re seeing though also with streaming is that there are tools out there that provide additional intelligence layers the buzzwords AI/ML, video quality analysis tools that do computational stuff around video quality like what Netflix does with their own standard and some of the other standards out that have been out there for a while. These are all becoming more essential pieces to prove out quality and availability and make sure things are going well.

Again, as we look at analytics today and particularly with streaming and live streaming, on the left I’ve got some examples of situations that exist today where essentially you have the inability to get at the data. It’s impossible. It’s a black box as we might note. It’s difficult otherwise either through technical means or other methods to get at that data. And in a lot of cases it’s incomplete. Missing data points, data from specific workflow components. My example there would be that there’s a lot of heavy focus today on CDN and player data. But I would argue that if you have a player problem, playback doesn’t work, that does not necessarily mean that the CDN or the player is the problem. The problem could definitely be what I call further upstream towards the source.

And so that’s what I refer to by incomplete. And sometimes the data useless. This may be a harsh word, but data needs to be helpful in solving problems, right? And sometimes you have data that isn’t helpful to fixing issues or to understand how well things went in a live stream. So this data needs to be more helpful if it needs to be actionable. Again, towards your business objectives. And a lot of folks again, if you measure ROI, look at KPIs, we’re all using streaming to try to increase our audience and decrease costs to as primary factors. Streaming is either a key part or a holistic part of people’s business and it needs to work and it needs to not be cost-prohibitive to use. And so these are kind of the objectives people look at as at the highest levels and why the data’s important to them.

So here, Wowza, we’ve been doing a lot of inspection on this. We have as you know, a server product and a service for live streaming. We have a hardware-based encoder today so we sit across the workflow and a lot of space components of that workflow. And really what people again are looking for is they’re trying to ensure that their streams are working correctly or not and good and bad is subjective. If you look at various data points you’re going to get different opinions across the customers that you ask. But we all have this notion of wanting things to work correctly according to what we’re trying to achieve. When things aren’t going well though, we need to have the ability to troubleshoot and resolve issues quickly, right?

So that ability to go in where an issue lies, know where that issue lies and go in and understand what the problem is and be able to resolve it quickly is very critical. As I said earlier, your event could be over by the time you resolve your issue or figure out where the problem lies. And again, just because you can’t play content does not mean the player for example, is the source of the issue necessarily. As we’re all looking at scale and quality and performance and reliability all come into play with that high quality and high performance, we often have large audiences. We often have the needs to have the same experience due to the widely distributed and large audiences. So we want to make sure that equally those things are possible for everyone that looks at our content.

And again, as I mentioned, AI/ML and VQA earlier, looking at making sure things are going well and when they’re not, can often be enhanced by further intelligence and not just log data per se. Log data is very key, but we also want people to know and understand and to be able to leverage more and more intelligence in the sense of automation and other pieces to optimize the viewer experience. And then greater interoperability and control include third-party integrations. This all comes into play again. Wowza as I mentioned, sits in pieces of the workflow. We don’t sit in across the entire workflow but looking across the entire workflow is super important. And that means that the more consistency or normalization you can have for data access, whether you’re using all Wowza products or you have a mix of Wowza products and non-Wowza products for example. Like we’ll talk more about and show with Fastly. Those are key pieces to make sure that you have as much as possible there to provide the insights to tell you when things are going well and not going well.

So, as we said earlier, ensuring live streams are working correctly is really key. The example I would give is again, with Fastly. Prior to working with Fastly our Wowza CDN component in Wowza Streaming Cloud — and while the CDN is a piece that you can use with Engine today as well, we were getting data at a much slower pace and rate for both analytics and for monitoring purposes versus Fastly where it’s near real time. And that’s a huge impact if your data is delayed and certainly longer than the length of your event, you may even not even know you have a problem until it’s too late. So that’s a real key to getting the data correctly or quickly and helping to understand if things are working correctly or not.

As we expand on timely troubleshooting and resolution for issues again, working with Fastly in this light helps us with understanding where there might be issues in their service impacting whether they’re for a single customer or for a broad set of customers. Nothing’s perfect, right? We’ll all admit that things could go down at any time. We work really hard not to have that happen, but when they do what you want to do is be able to resolve these things quickly. Part of that is understanding general consumption too. And regions definitely play a role here and understanding that so… With scale and reliability, and again, quality and performance.

And Jim chime in here if you want to.

Jim Hall:

Yeah, I was going to say, so Fastly’s real-time log streaming is somewhat different than some of the other folks who do provide real-time log streaming in that it covers all of our network. It’s not a segment of our network. It’s not just web traffic. It’s everything. It’s TLS, HTTP, video, web. It doesn’t matter. It’s all one network. It’s not just a sub-net or traffic type. We also deliver the data for every request, not just a sample of your data. We can sample the data if you prefer. But, most people want everything and they want it in real time, not minutes or hours later. And the other is that our log stream scales massively. Taboola gave a talk at Fastly’s Altitude about how they’re leveraging log streaming to deliver over a million lines per second. So this is real time 100% of the network and it scales.

Jamie Sherry:

Thanks Jim. As I said, greater intelligence and automation can help in optimizing the viewer experience and the success of your events and live streaming and streaming as a whole. You know, log files, again, really key. But data going mainstream means leveling up. And so again, Audio Visual Quality Analysis, competition analysis, AI/ML looking across hardware and software and networking applications is really key. And I have these four words up here, inspect, interpret, predict, and prevent. I like to use these words to talk about not only what Wowza Insights will provide for customers, but what really is I think, key to providing good monitoring and analytics in a modern data world. Which is not only just looking and giving visibility and observability for the data and giving that across the workflow.

But being able to interpret things right, as I said, being able to tell you as a customer where a problem lies, and even perhaps what you should do to resolve it, is super helpful. And time is money again, right? And so that helps shorten that time. Now predict takes that even further. Look at trend analysis and you could look at situations like say your bandwidth for your encoder coming into your origin or ingest infrastructure. And if you start to see fluctuations there or decrease or something else and what a user would consider a negative or a sorry, a customer would consider it a negative situation, perhaps you can prevent that situation from being a catastrophic situation.

And you know, prevent again, takes that kind of predictability even further by suggesting that actions could be automated. Imagine that you define a data point that hits a certain threshold. And on that threshold you want an action to be taken, to recover from a situation or to improve a situation. That’s all possible today with the right pieces in place to help further mitigate downtime and even prevent just a negative experience with users.

Touched on this area again, just to dive a little deeper greater interoperability and control including third-party integrations. So again, as I said Wowza sits across a majority of the workflow for live streaming, but not everything, right? We don’t do everything. And, at the same time, the flexibility to be able to put together a workflow that includes a mix of Wowza a non-Wowza is super important. And the ability with great ease and feasibility to build and run those applications, get that reliability and quality performance, those are key pieces, right? But to look at the data around that to make sure that it’s actually high quality, high performing and highly reliable is super important. Again, as Jim said, Fastly plays a key role in that kind of situation.

There are in a lot of configuration log data the ability to tie the data flowing through the CDN and back to the player consumption back to the encoder all the way through the workflow is super important. And, and that data can flow through all kinds of QoE platforms and services today like Wowza’s. But also folks in the industry that are doing similar things like Mux, Nice People At Work with their platform and a Bitmovin as well. And there’s others in the industry that are doing similar things. Key is to be able to tie that data together and give some intelligence to that and give meaning to that so that, again in that sense of where it might look like a player problem, but it’s really not a player problem that you can go narrow down where that situation or issue really lies and resolve it so.

Demo time.

Justin Miller:

All right, well that being said, I am going to stop sharing my screen and if you’ll give me a moment, Jamie is going to get on and show you a demo so.

Jamie Sherry:

I’m going to give it a try. You know how demos go?

Justin Miller:

I do.

Jamie Sherry:

Alright, share my screen. Okay. We can see my screen. Okay. So I am logged into Wowza Streaming Cloud service. We have a bunch of sample streams that are up and running all the time for various reasons. And so, just to show you what this content is. You got a player. Play this back real quick. So we’ve got a live feed out of our Berlin office. It’s getting to be nighttime there but we can still see some of the ladies out there, which is great. So there’s that stream, the live stream. Now this set up is running through a ClearCaster, a camera into, in the office, into a Wowza ClearCaster encoder, a hardware-based in we have.

And it’s then going into our origin and transporting infrastructure in Wowza Streaming Cloud. And then it’s going into Wowza/Fastly as a CDN. And then out to a player, in this case, as is with most content, as a playback piece. What I’m going to do here now, I’m the only one probably watching this at the moment. What I’m going to do is introduce a Load Test and an audience to join me in watching this great content.

And now this data that’s coming out of this workflow all the way across the workflow from encoder to the Wowza Streaming Cloud origin and transcoders to Wowza Player. And in the middle I skipped over at Fastly. That data is all flowing into a Wowza Insights and our centralized storage collection and processing infrastructure for data, real time. What I’m going to show you now is a dashboard that is a conceptual piece for what you could do with that data. What you could realize with that data. So we call this the real-time dashboard and again this is a demo proof of concept piece.

Let’s check load test here. We’re still provisioning that. We should have started that a little sooner now that test is adding those users and then we can see now we are at 52 users. So, two pieces I’m going to show you: This dashboard has a bunch of what I call a pulse check. Quick items up top for stakeholders that want shiny quick answers on things. We have a viewer count. We have a stream score which is actually not working for the moment. It should pop in there though. We measure the latency across that workflow. That’ll show up and that’s available. There we go. The score is up. Time to first frame for playback. We calculate that for all the players watching as well.

Those numbers should start showing up here. As I go, down below we’ve got our workflow and we’ve got again as I said, a ClearCaster going into Cloud and our origins where we have a single stream going in in three renditions, created an adaptive bitrate set going through Fastly and then out to playback. The green bars up here indicate that the status up here is green just to let you know. So now we’ve got a time to first frame numbers showing up. Various metrics showing up in here. This dashboard shows you tabula data versus little more of pretty data with visualized the data, we’ve got map down here, which is showing you the full path of that from source to playback all the way through. As I’m going to show you.

We got a preview over here. All this data is being updated. It every 90 seconds, these little trend lines show you how things are going. As I mentioned, trend analysis is important to mitigate or prevent predict issues. Over here we’ve got a bunch of data over here as well because it separates up by these different columns. This graph here, however, and it’s because I just started this that it’s not showing the full last hour, but we’ll show you the last hour again, speaking of trend analysis yet kind of a better indication of how things are going with certain sets of these data points here.

So, this is again what people could do with this data coming out of Wowza Insights as an idea this could end up in our products going forward to in the second half of year, we can touch base with you, keep tabs with you on that kind of thing. This is a Cloud-specific workflow. We’re also working on an Engine-specific workflow as well. So imagine if you’re an Engine customer that you could do this as well with your origins for Engine. And then go into Fastly and go out through a player that way too. So what I want to share with you over here though, which is interesting, I should have showed you the before and after. I’m going to log in.

Justin Miller:

And that being said, if anybody does have any questions, please feel free to enter them into the Q&A window. We’d also love to know what kind of analytics you’d like to be seeing, if you want to enter that into the Q&A window as well that we store that information.

Jamie Sherry:

So I can get more of a real-time set of our usage data in cloud as an example of speed and processing time for the data. I believe that we now get this data from the time we query it, get it from Fastly to the time that it’s in our infrastructure is down to two seconds. So I can consider this essentially real time. If you go back to the dashboard, we can see all these playback regions over here and the counts for the player. And I have an instant mapping of that over here as well now. And this was all instantaneous, which is super helpful from a usage perspective.

So we’ve got the dashboard again that shows a visualization piece for customers who would prefer something like this. You have certain operations folks that might be a benefit from this and this does look at a single stream and you could imagine a multi-stream view where you can drill down as well. And certain stakeholders at higher levels might want the warm and fuzzy, like I said, a quick pulse check at the top or the kind of pretty visualization is creates but very helpful in terms of information as well. There was lots of good data in here at each component level to help. This dashboard, actually I can show it at the moment but actually can invoke alerting too, we have actually a section up here at the top that could show you alerts for situations where you have to find predefined data points to hit certain thresholds and you want to be notified when those happen.

So imagine that you could have an alerting notification concept in here as well. But again, in Cloud, what’s realized today with our integrational facet, which is GA today, it’s been out for a month or two months almost, is this ability for you to use this data show up in real time.

Justin Miller:

Jamie, let’s put up the real-time dashboard again if you could.

Jamie Sherry:

Sure.

Justin Miller:

And I do have a couple questions coming in. One of the things I just thought we’d answer them right now. I have a question coming in about is there two costs for Wowza CDN and Fastly or is Fastly now the Wowza CDN?

Jamie Sherry:

So Fastly, that’s a good question. In full transparency, the Wowza CDN product, which works for the engine is not yet production-ready with Fastly. It still uses our current CDN offering. That is a roadmap item for this first… well I believe it’s the summer timeframe. So we can come back you guys, anyone who wants to know definitively those timelines. The Fastly integration in our Cloud service though is a part of that service natively and it does have a cost component in Cloud. But if you look at our Cloud pricing you have a delivery cost component with the cloud service regardless. So it’s part of that cost structure there.

Justin Miller:

All right, great. And I’m also being asked, is this dashboard on Wowza Cloud? I’m not sure if it’s this dashboard we’re looking at right now in this question or the previous one. Is this dashboard on Wowza Streaming Cloud free to access or is it in chargeable service? We’re using Wowza CDN for content delivery.

Jamie Sherry:

So this is a proof of concept. This doesn’t exist in any product today yet. The intent, what we’ve been doing as a… and I can talk a bit of roadmap on Wowza Insights. So we are working towards building out in the next quarter an API to this data and that will include Engine data and Cloud data. So and ClearClaster data, in fact. So every Wowza product data will flow into those insights. You’ll have a public API to get that data out and do something with it. This dashboard makes use of a private version of that API today that will again become public and in the summer timeframe. And this dashboard could end up in our cloud service as a visualization piece. It could end up in other or Wowza products and offerings down the road.

As we work more on solution selling with customers, imagine the ability to have an insights UI for people holistically for their workflow where you describe your workflow visually and then Wowza Insights in the sense of a dashboard like this sits on top of that workflow, picks that up and just displays that workflow for you today. Again, this is a fixed workflow up in the four boxes there that you can see. But imagine you could define your own workflow with a mix of Wowza products and non-Wowza products and then the dashboard just shows up and it shows you everything.

That’s the vision that we see, but we’re starting with an API and then we will work on visualization and user interface tools. As, we get past that API we’ve been doing a lot of discovery and collecting feedback from customers and others on what they want, what they want first and second. That’s why we’re working on the API. It makes sense to build UIs on top of APIs. So we’re working on API as I said, second quarter, all of you, if you have feedback on what you would like in terms of visualization or the API or other pieces, we are envisioning a lot of other pieces.

One thing we had in the presentation that I didn’t touch on tremendously is this notion of looking across the content life cycle. Live content has a life cycle to it, right? And that includes the ability to look at your live events before you go live during while you’re live, which is like what we’re seeing here, and after you’re live, imagine this tool turns into a post engagement piece where you can go back and look through your live events and see where you can improve for the next time. Things like that are what I described as the content life cycle. And those are pieces that are coming down the road too. And again, we’d love your feedback on what’s important to you, that helps us drive forward on the pieces we want first to provide and the second and so forth.

Justin Miller:

All right. Thank you. So Alex and Miro, I hope that answers your question about getting data through the API. If you do have any suggestions as Jamie mentioned, please go ahead and also type them into the Q&A window just so we can store them somewhere. That way Jamie also has access to look at those suggestions later. We have a question from Harick. Does this as well support concurrent streams?

Jamie Sherry:

So again, this dashboard being what it is, is looking at a single stream. We envision with any user interface pieces that we do having a sort of multi-stream approach where you would have simultaneous events running and you want to have a heads up display or a dashboard of those with key metrics that you can look at it. And again a quick pulse check kind of approach. Like the data at the top of this dashboard and you can drill down into a single event, see things on that multi-stream view that help you understand whether it’s going well or not. But then you can, if you really need to drill in and say,” I have a question about this one event or I want to see what’s going on, or there’s a problem” and you can drill down so.

Justin Miller:

And I can see some people really like these dashboards right now. Kumar is actually asking, can I create multiple-tenant dashboards for each customer using the Wowza Streaming Cloud encoder?

Jamie Sherry:

So again, the idea would be… then again this is a proof of concept. The idea in future UI work would be, it depends if you’re going to use cloud and kind of a white-label customer-to-customer approach. A more of a B2B approach, which we do have a lot of customers doing right there to build their own UIs. You can imagine building dashboards for customers or letting them build one. Build them too, or even having them be available to the customers for their events so that that kind of B2B approach would be possible. If I understand you right.

The other concept that I didn’t touch on yet but I’ll touch on now is while we see people that want visualization to that point, they may want visualization in their own interfaces, their own management interfaces. So imagine you take the Cloud API as an example and you go build your own user interface for managing live events. And part of that you want is smaller analytics. You could imagine you could embed dashboards in there so you could create dashboards and then have them show up in other tools down the road as in the sense of broad analytics today, embedding dashboards is becoming a more popular thing where people don’t… there’s this kind of middle ground, right?

Where they don’t want to build them themselves. They don’t want you to do everything and be locked into your interface. They want what you have, but in their own interface. And that’s what we call them middle dashboards.

Justin Miller:

All right, thank you. It sounds like though all of this is being pulled into the API, so they could either build it themselves or maybe even contact a professional services to build something for them.

Jamie Sherry:

Correct. I would say if you have opportunities where you want to look at this data and talk about visualization now for Engine or Cloud in particular, our Professional Services group is… while we’re building out this API in this second quarter and we’ll do more right behind that with all the other pieces I mentioned. We have a Professional Services group, they are working with customers right now on Wowza Insights opportunities where they will leverage that API. And we’ll be looking at embeddable dashboards, visualizations, or they will go build them themselves. So our team would love to talk to you further about such opportunities.

Justin Miller:

Now I think you’ve already stated this on the roadmap that given a timeframe when these dashboards might be available in Cloud?

Jamie Sherry:

So the availability of visualization is further out. I would call it earliest, second half of this year.

Justin Miller:

Do you think that’s going to happen before these options for is it Wowza CDN with Fastly are available as part of Engine?

Jamie Sherry:

Has to time, right? If it doesn’t, well with Cloud it’s okay with Engine there is a question of whether these things would show up in the user interface of Engine. I will be honest, that may or may not happen. They may end up being insights as a whole may become its own interface that people will access and still be able to use with Engine, just not through the Engine Manager, user interface. And the tie in for anything that leverages this. We’ll have to coordinate and time with the Wowza CDN for E          ngine using Fastly.

Jamie Sherry:

Obviously if you can’t get the data in real time, there’s no point of visualizing it.

Justin Miller:

I’m not actually sure if they were talking about the dashboard for engine or just specifically connecting Wowza CDN with Fastly. The API connections we’re talking about.

Jamie Sherry:

Oh, well I’m talking about a broad API for getting data out for real-time monitoring analytics. The notion of integration of Fastly with Wowza CDN love to come back to you on a timeframe. I believe that the earliest it could be as a summer, but again, getting it in Wowza CDN for those Engine customers versus what they have today is something I have to confirm.

Justin Miller:

I’m being asked by one person, does this require the use of Wowza Player?

Jamie Sherry:

It does not. So I have Wowza Player in here for the moment just because that was the quickest and easiest way to demonstrate a preview playback in here. But the intention down the road is that any visualization like this would have the ability to use any third-party player. We are working with third-party player vendors right now to integrate data back into our infrastructure. So that you have that fourth box on the right that isn’t Wowza Player and you get that client-side data in real time.

Justin Miller:

Right. So as we can access more data, we have more ways of placing it.

Jamie Sherry:

Yeah. In a third-party sense, regardless of where you talk in the workflow, the idea is to try to get as much data as possible. With Wowza products in that box, you’ll get more today. But we realized that you’re going to have a workflow most likely as a customer that is a mix Wowza products. We don’t provide everything anyway. And so we want to make sure that you’re getting as much data as possible for the purpose of monitoring analytics so…

Justin Miller:

Now I have a few questions that came in that I’m not sure if they’re related specifically to just questions on types of analytics they can get here, but I’m just going to go through these. Noel says, how did you do that load testing? Is it something that’s generally available within Wowza?

Jamie Sherry:

It is not available today. We did pilot this with a few of our customers last year, which helped inform on some of the things we’re going to do in Wowza Insights in this calendar year. So not available today, but intended to be available as part of some of the features we’d want to do in Wowza Insights going forward.

Consider it second half of the year at the earliest. The specifics I would say again is imagine the ability to test your event with a simulated audience before you go live and then get out a report card to tell you how it went. Including suggestive areas for improvement. You had a bitrate/codec mismatch, or a bitrate/keyframe interval in this mismatch, something related to the streaming. Something related to a region that couldn’t get out the content. In an ideal way. You’ll get indicators for how to improve it out of running such a test. So yes, again, imagine that load test piece becoming part of our offerings going forward, but just not yet.

Justin Miller:

Another question is, could use this to define why viewers are facing buffering issues in a particular region?

Jamie Sherry:

Yes. The intent on the road is to provide as much context in the specific parts of the workflow as possible when there is something that you define as a customer is not going well. So again, I like to say that various data points that are measured in real time from our analytics are subjective, right? You might say that, Hey, this cache hit rate of 68% isn’t good enough for me. I want to do something different, something better, right? Or it is good. And that’s not Fastly by the way, the reason my cache rate is where it is, is just because of the nature of how to run this test.

So usually those numbers are up in the 99%. So the situation, the answer’s yes. The intent is to cross the workflow, provide as much information as possible to tell you where we think something could be better. Something is wrong, something is right. And again, as definable by you with all of this data, because again, it’s subjective.

Justin Miller:

All right. And by the way, I’m just rereading these questions here. Some of them I know are repeated parts. Alex was asking, can we have a finer detail of time segments? Some of the events we stream or maybe only two hours long and shorter segments would be useful to see your patterns. Love the dashboard. Can we have that in this time soon? Can we show captured info?

Jamie Sherry:

So a couple of questions in there. I’m not sure what is meant by shorter segments unless we’re talking about the analytics I showed in Cloud.

Justin Miller:

And feel free to respond Alex.

Jamie Sherry:

Alex I just might want some clarity on what you mean by shorter segments. Again, all this data in here is updated much quicker than in the two-hour window.

Justin Miller:

Yeah, actually he is asking if you could share the dashboard that we see today.

Jamie Sherry:

So going back to Wowza Streaming Cloud, so all of this when you use Fastly is, is real time is as the data is coming in.

Justin Miller:

Once the API is done, so we have a place to start. So he wants to know how the dashboard might change. I think once the API works with Fastly to have more real time.

Jamie Sherry:

So when we have the AP,I to be honest, what we will have is people go and build and grab the data and build their own dashboards and their own tools first. When we have the ability to put a dashboard like this in our product or into an insights user interface, it’ll be customizable. Like I said, you’ll be able to define workflow, define elements within reason, right? We’re not going to be able to have a complete builder that lets you do a million things, but we’ll reflect on the things that resonate well in here with customers and if they have suggestions on other pieces they want to see and make those available as well.

The key is about the data and the API will again further people’s initiatives that want the data but want to do the visualization themselves. That’s the first part. And again, we realize that these realizations are certainly helpful for people that don’t want to build them themselves. And, those pieces will come later with the ability to customize it at some levels.

Justin Miller:

So Alex I hope that helps. I’m not sure what you’re talking about having a place to start. Are you trying to figure out what API data is going to be available to you, and are you asking that in relation to what Cloud data that you currently can see, I think relating to the health monitoring, perhaps. If you want to give us some more detail, certainly we’ll be happy to help answer your questions here.

If you want to give us some more detail, certainly we’ll be happy to help answer your questions here. I’m just going to jump to some other questions. I was asked by a couple people doesn’t really relate to our analytics per se, but I guess they’re asking if a Fastly is going to help us with DVR capabilities or a VOD streaming or any of that with Cloud?

Jamie Sherry:

Certainly there’s a roadmap now that we have Fastly in place and I’m not working on it anymore. I used to work on it, but I envisioned the roadmap this way before and I think it’s still the case. We’ll have to get back to you on any timing. But the integration of Fastly has opened the door for a lot of new enhancements and new features — one of which is DVR. So we envisioned DVR showing up in Cloud at some point.

Jamie Sherry:

Timing again is kind of TBD. I’d have to get back to the product manager who’s running with Cloud now and have them respond on such things. But DVR, the way we do our architecture with Fastly now creates a whole additional set of lots of VOD capabilities and really content manipulation capabilities in terms of things like clip extraction and manipulating content that could be targeted at specific users for specific reasons. There’s all kinds of scenarios where the doors open with the new architecture and workflow of Fastly.

Justin Miller:

Awesome. Glad to hear it. So it’s more possible than it was before. Definitely something on the horizon.

Jamie Sherry:

Don’t have the dates. We have to consult the Product Manager for Cloud, Robert to get those answers.

Justin Miller:

And just to respond Alex was the one who was asking relating to what we currently can see within Wowza Cloud. He says, what I see here would be a huge start for small companies with limited development time.

Jamie Sherry:

Yes. We recognize visualization definitely has an audience. And so imagine places like this tab in cloud under the transcoder with the health tab having additional information here, more like what you see with the real-time dashboard.

Justin Miller:

And by the way, the person previously was saying we are so waiting for was a VOD with Cloud. So they’re excited…

Jamie Sherry:

It’s coming. Thanks again Robert the Cloud Product Manager working on it. So it’s coming.

Justin Miller:

A few last questions here and I’m not really sure if this one even relates to anything doing, but I’m going to go to these here. So Harkin was asking, does this dashboard also help troubleshoot issues? I mean from where the issue is, let’s say from source or Engine or with Cloud or maybe with the player side?

Jamie Sherry:

I mean, the idea is to present in this dashboard, primarily present… Let me see if I can trigger an alert here. We have this sneaky little thing down here that can introduce errors into the test blow test. The dashboard’s designed today is to provide information that’s configured and measured. So there’s a mix of things here that are configured and measured, for example, in the encoder, this information over here, the frame size of the frame rate is configured. But what’s measured is… well a lot of this is configured right? And then what you see are some measure pieces across the board here in different places.

But you see it makes it configured and measured is really the overarching message there. And the idea here is that by looking at this data you’ll have some insight as to yourself in your head mentally that you know what’s right and what’s wrong, right? If you see something crop up that doesn’t fit a measure value you would expect or you look in here and you say, Wait a minute, I thought I configured that frame test to be 1920 by 1080 yet it doesn’t say that I need to go back and check something.

So, it’s really more of a kind of a visual check and verification that way. Now again, what we see over time is the ability to inform on things with a rules-engine approach that ties in alerting and thresholds where you say, okay a rule could be like, if this combination of these two data points is this far off or this mismatched or something of that nature you could create these kinds of situations where you could be more intelligently notified. For example, if you saw a region in here you were trying to block based on a geo rule and somehow that got through the map might be a great way to tell you, “Hey, why are people watching in Hong Kong? I clearly set up a geo rule with Wowza and Fastly to say no one’s allowed to watch in Hong Kong.”

And then you could be warned on that and you could even have actions taken, right? You could decide, hey, if somebody slips through the cracks, shut them off. If somebody slips through the cracks, go check and make sure that because we all know about IP addresses in geo resolution, it’s potential to be error prone, right? That you could actually have an approved yeah, they’re okay because they are really this and I know that and that’s okay kind of thing, right? But it can all be automated or rules-based so… again, the dashboard just shows you information that you kind of meant to interpret and decide for yourself. It does not look like my rules are going to cause an alert to show up here.

I’m usually messing with the rebuffer and failure rates, makes these numbers over here go up. But that’s how demos go. Sometimes they don’t work. So again, long, long answer but that’s how, that’s how this dashboard works today.

Justin Miller:

And we’ve got about two more questions left before we go. I just want to say to Jim, thanks for joining us. I feel like you haven’t had a chance to say much.

Jim Hall:

I was just going to say when one of the what a great monitoring and alerting model for live streaming is, depending on your workflow, 304 monitoring, if you suddenly see a spike in 304s, it means that the manifest isn’t updating and it means something might be going wrong between the encoder and the packager.

Justin Miller:

Yup. All right. That’s a great tip. Yeah. Thanks Jim. Anything else Jim? I wanna throw out there?

Jim Hall:

No, I think just that one of the awesome things about the integration with Fastly is how flexible it is and the data points that you can add custom values, et cetera. So you can really make this work for you. As opposed to just a fixed set of data points that you just have to work with.

Justin Miller:

Yeah. All right.

Jim Hall:

Yeah, we are super happy with the Fastly relationship. The implementations did not take very long. We have a good smart team who’s eager to work through all that. But you know, really happy with what we’re doing with Fastly. Our real goal now is to get that traffic growing and moving so lots of benefits.

Justin Miller:

And definitely let people are hoping for that VOD coming up.

Jim Hall:

It’s coming, Robert knows the timelines.

Justin Miller:

The last two questions I have I think are more feature requests. I’ll just mention them. Mark was saying will it be a standard function to export the analytics they could give a customer to archive for later?

Jim Hall:

Yeah. So one of the other features we envisioned is what we call a raw data export. So imagine you set a time window or you define it by the name of the event or stream. And then you instantly have the ability to download all the data that we provide on that event and send that off so…

Justin Miller:

And Crystal’s saying if blocker for us is not having a way to restrict the privacy of our streams to a domain. Is that a feature in the works in the future?

Jamie Sherry:

So that’s a Cloud- or an Engine-specific thing. I would have to defer to our product managers on those parts.

I mean, I know you can craft stuff with engine with cloud. I forget if that features in cloud. I don’t believe it is today. So best to confirm with, with the respective program managers.

Justin Miller:

I think Crystal, you may just want to contact our support for help for things like this.

Jamie Sherry:

They’ll know right away.

Justin Miller:

Yeah. Last question from Noel, she said we lost some health info with recent changes. I’m thinking of GPU percentage will that be returning?

Jamie Sherry:

So I don’t know where that would… I probably need more info to know if that’s…

Justin Miller:

Yeah, we need a little more info. To get to understand it, to be able to troubleshoot things isn’t really what this webinar is for. So definitely appreciate you asking. But certainly our support would be able to directly figure that stuff out much more quickly than he could right here. Noel was saying it’s cloud, but again our cloud support team definitely can help you out with that.

Yep. All right. Well, Hey, I want to thank everyone for joining us again, Jim, glad to have you here. I’m sorry we didn’t get a chance to chat more with you.

Jim Hall:

That’s okay.

Justin Miller:

All right. And Jamie, thank you for coming as well. It’s definitely a great demo. I love this real-time dashboard. I’m sure many people are excited to see something like this put into effect. If you want to see it faster just email back to us at webinars@wowza.com and let us know what you’re interested in seeing. We definitely want to be able to give people what they want. Alright. Take care everyone and have a great evening. Thank you everyone.

Jamie Sherry:

Bye everybody.

Jim Hall:

Thanks.