Eric Kavanagh: All righty, ladies and gentlemen, it is Wednesday, June 21st. It’s 4:00 Eastern time and of course that means in the world of enterprise technology it’s time for Hot Technologies! Yes, indeed. My name is Eric Kavanagh, I will be your host and moderator for today’s event. It’s a hot topic folks, it’s a big one: “En Marche! Enabling the Mobile Workforce.” And I did not intentionally grab the tagline from Mr. Macron’s candidacy in France. It was quite coincidental, I promise you, but it’s still rather exciting. So, we’ll talk all about the mobile workforce and how you can make sure that those folks get what they need, and they can do what they do well. Lots of challenges, lots of issues out there. We will archive this webcast for later viewing, so if you miss something you can come back and check it out. Also, share it with your friends and colleagues.
And I should say do not be shy; the best way to get really custom content and the information you need from an event like this is to ask questions. So, you can ask a question from the chat window, or from the Q&A component of your webcast console. Anytime during the event, send it on and I will be sure to grab that and weave it into the Q&A at the end. We’re going to have a couple of presentations and then we’ll hear from Bill Ellis of IDERA Software. Of course, our own Robin Bloor is on the line today. And with that, let’s dive right in.
So, I’ve got some good stats from RCR Wireless about what’s going on, and really it’s pretty mind blowing. They say the global mobile workforce will hit 1.87 billion people by 2022. That’s over 40 percent of the overall workforce on the planet. So, if you think about that, now, all of the sudden where you used to have, in terms of IT capabilities, in terms of functionality on devices like computers, where you used to have 99 percent or more of that on premises in your offices – that was even, let’s say 15 years ago, 10 years ago it was probably 85-90 percent, five years ago it was like 70 percent? Something like that? Now, it’s all the way down, almost to 60 percent. And this is a big deal. So, we’ve seen this massive shift in terms of the technology, the actual tools that people use move outside the office, into the workforce.
Well, there are countless benefits to this. I mean, literally if you look at the shipping industry for example, like UPS, or if you look at guys who go out to the rigs in oilfields, if you look at any of the various jobs where it helps to have deep functionality with you, on the road, the mobile workforce is changing everything. Now, one of the problems – and we’ll talk about this is some greater depth – is that we have a couple of different things going on, one of which is diversity of workforce. So in 2020 – I just saw the stats today – there are going to be five generations of people in the workforce. That means you’re going to have grandma and grandpa and then mom and dad and also the kids, but theoretically you’re going to have essentially great-grandpa and great-great-grandpa and great-great-grandma out there. Now, obviously it’s not within a particular family, but the point is generational-wise, you’ve got five different categories of broad individuals in the workforce, each of them has their own tendencies, their own predilections, their own propensity to work with technology.
Obviously, kids tend to be mobile first in terms of how they interact with the world. And just think about the communication channels that it changed – we talked about this on another show recently; SnapChat is how a lot of teenagers communicate, they don’t really even want to talk to you on the phone, they just want to send little SnapChat messages back and forth. That’s just one example in the consumer world of how things are changing, and that could be spread out across the entire spectrum of technologies, of functionality, of individual, of company, of business model. It’s just all over the map, but the point is that the mobile workforce is real, it’s out there and unless your company has a solid program for understanding how that affects your business processes – and I’m talking very specific technology-driven data-fueled processes – if you don’t understand what those are and are not managing that through an IT infrastructure and a process and a governance perspective, you’re going to have all kind of problems.
So, there is the iPhone. I remember when that sucker came out, it seems like a million years ago now. But it was only like what, 2007 or ‘08? It was not that long ago that we didn’t have iPhones, and of course the form factor just fundamentally changed technology, and really enabled the mobile workforce. And I recall of course at the time, the iPad came out and then the iPhone, right around the same time. I can’t remember which was first, but the iPad was really one of the most significant forces of change for enterprise IT, possibly since the mainframe. And the reason is because frankly, a lot of very senior executives, C-suite people of large organizations loved it right out of the bat. And said, “I want it. I’m bringing it to work.” Well, think about that – all of the sudden IT had to turn around and deal with the problem that they probably didn’t want to deal with, which was deal with all these new devices.
So, now, if you had iPads – well, how do you weave that into the matrix? How do you maintain governance over that? These are all really big challenges and that old iPad and the iPhone really was a hugely disruptive force in IT and in IT management for a lot of organizations, large and small. So, we still have this spectrum of challenges and benefits that ranges across about as broad an array as you can imagine, with mobile devices. And of course, they keep changing, right? So, now, it’s not only BYOD, it’s BYOA a lot of times, where executives and professionals are bringing their own device. Well, we used to call that “shadow IT,” right? For those of you in that older generation, you may remember the old radio shows, they had radio drama and one of them was The Shadow – “Who knows what evil lurks in the hearts of men? The Shadow knows.” And I remember that because I was a kid. Well, shadow IT is fricking everywhere these days; everyone’s doing shadow IT.
So, this is a real challenge for the IT management and business process management, all the operations people. You want to be able to leverage mobile devices, but you want to be able to tie that back to your systems, and there are lots of weird, little issues that come into play. Not the least of which is the visual experience and the associated functionality that you get when you use a mobile device. And any of you who have used multiple devices like an iPad, versus a laptop, versus a desktop, versus some of the newer mobile smartphones that are coming out, having experienced the fact that the functionality doesn’t work quite right, and this is a real problem. In fact, the browser wars should have prepared us for this, because the browsers all do things slightly differently, as well. And that’s another big challenge for not just design, not just the look and feel and the sleek nature of the application you’re using, but the actual functionality. How do you get the drop-down menu to select what you want on that device? That’s a big deal.
So, that’s what we’re going to talk about a bit today, and we’re going to hear from Robin and Bill Ellis, as I mentioned, who’s a real expert in this field. So, this is one of the big issues that people have – it’s just that the darn variety and there’s no single method for being able to work across platforms. You’ve got Samsung and Apple mostly making these things, but there are all kinds – there are so many devices! I saw recently that the iPhone was winning in terms of sales, and I was shocked at how low the number was – it was like, I don’t think it was even 20 percent! And they were number one, which means there are literally scores – if not hundreds – of devices out there that can be used. Well, you can just imagine how the IT department feels about that, and of course, that range of technologies is changing; it’s getting more diverse by the day.
Everything is changing, we’ve got all kinds of stuff happening – containers, just to throw another wrench in the works here. And then, of course, we have the diversity of the workforce. A lot of millennials, they’re just a lot different in terms of their preferences, how they use technology, what they’re willing to wade through, how quickly they can figure stuff out. Typically, it’s faster than with us old timers, but nonetheless all that has to be mapped back to your on-prem systems, or at least up to the cloud. And that’s a big, big challenge.
And with that, I’m going to hand it off to the inimitable Dr. Robin Bloor. Robin, take it away.
Robin Bloor: OK, thanks for that brief introduction. Let’s talk about mobile. It wasn’t particularly obvious – Eric’s referred to the introduction of the iPhone – it wasn’t particularly obvious when the iPhone came in exactly what this heralded. I think it became obvious when the iPad entered in that we were actually going to have a fairly diverse mobile world. I’m a kind of Apple bigot, really, so I don’t really think in terms of Android, but of course, although Apple makes the majority by a long way, the major profit from both the pad market and from the phone market, it doesn’t have the numbers anymore, which is kind of an interesting thing. And that means that there’ll be – apart from anything else – there’s going to be new devices, people are going to take them up and they’re going to sell in the millions. So, it creates a very diverse environment, that you may need to go through.
The joke here of “I’d ask Siri where the hell we are if I could get a signal.” The thing that makes mobile devices slightly different is that desktops are connected all the time. And mobile devices are not necessarily connected and they’re not necessarily 24/7 on, ’cause people can turn them off. also you can get them to airplanes and things like that, and therefore it’s a different kind of device than anything you ever had before. I would maintain that the mobile phone is actually the real personal computer, ’cause it’s the one that you have with you all the time. It’s the defining human mobile device. The tablet is slightly different; it’s a kind of weird situation, that when you think about it, that in one way or another, there’s more than one functional kind of mobile device.
Anyway, what it means to be mobile. The internet changed. We didn’t kind of notice it happening – I didn’t notice it happening – but nowadays 80 percent of internet activity is from mobile devices, and that’s an extraordinary figure when you think about it. But 47 percent of that 80 percent is tablet traffic. It’s possible to provide most applications in a mobile setting. In other words, if you’ve got applications that already exist and, you know, they’re accessible on the desktop, you can probably put them on a mobile phone, but there is obviously limiting factors. Form factor and keyboard are one of them. Tablets themselves are, according Microsoft and Apple both, are gradually going to replace mobile PCs. And they have particular applications in certain areas, because they’re more rugged.
One of the things that I remember talking to healthcare IT people about, was the fact that before the tablet existed, if you went into an environment that was an isolation ward, you know, you would have to have your devices that you took in with you, would actually have to be disinfected in some way or other. It’s really easy to do that with a tablet, it’s not at all easy to do that with what they used to have, which was desktops that were mobile by virtue of being on a trolley and plugged in in the environment. They used to have to kind of stay in those kind of environments, or go through an extraordinary kind of disinfection being taken out of those environments. And we don’t kind of think much about those environments, unless we work in those environments. But the tablets and the mobile phones have made working in those environments really quite natural to be connected and work in those environments.
And when the stat that Eric put at 1.7 billion, I think it was, mobile workers by 2020. Am I a mobile worker? I kind of think like that it’s, I am a mobile worker in the sense that occasionally I am working outside the office and when I do that, I’ll be working on a tablet or doing stuff on a mobile phone. So, when you actually look at that, and you think about that, it’s probably on account of people that will only be using mobile devices for their workforce, so people that are actually fundamentally moving around. Anyway, you can think now in terms of three kinds of user: desktop users, tablet users and phone users. And they need different applications. And that’s the reason for mentioning it.
Camera and voice are now an inherent part of mobile devices but they’re also an inherent part of desktops. But they’re used in different ways on mobile devices and they have different interfaces on mobile devices. And the whole character of why you’re using that is different on a mobile device. So, it’s if you’re building mobile applications, you are not building the kind of applications that you used to build, for a whole host of reasons – a lot of which were on that slide. So, if you were a business that already in one way or another was building applications that ran on websites, the question is, should they be mobile applications as well? And this slide kind of looks at that. A web application, you can do more on it, simply because they’re built in one way or another, they’re built without actually caring about the form factor, so people will build a webpage which you can’t reasonably use, or you can’t easily use on an iPhone or an Android device, which might just be usable on a tablet, but even on a tablet might not be particularly good. Normally, it would be OK.
Or you can build a mobile app. If you build mobile apps, then there’s an application glut on various download stores, and that kind of drives down their resistance. If you look at my particular iPhone, it’s just packed full of applications that I can’t seem to get rid of; I delete them, but they always seem to get downloaded again in some strange way. I obviously don’t know how to manage an iPhone properly. But you know, you end up with just a glut of applications and it doesn’t make any sense. I got more, I suspect I’ve got more applications on my iPhone than I’ve got on my desktop, which is bizarre when you think about it. Mobile apps are a litmus test for success. It’s kind of interesting that some web businesses – Yelp is one of them – did extremely well by creating an app and having people download it. And it seems that the areas where there was reasonably good success were actually in the financial sector; that’s banks but also E-Trade and companies like that, ’cause people like to be able to trade stuff on the move, at times. Food applications, so not just looking for restaurants but also making recipe sites, they did really, really well in terms of apps.
And a lot of people didn’t do particularly well at all, and that’s the reason, I think mostly is that there’s only so many apps that you ever get used to using, and if you’re only using an app once every few days or so, then you forget about it. If it doesn’t have a big personal value to you, you kind of forget about it. So, it’s difficult to create a mobile app that’s accessible in the general sense, but obviously you can create them for your own staff and use them within the organization. Mobile apps have really big development costs, and it’s a number of reasons for that. One of the reasons for that is you’re actually pointing at a distinctly different number of devices.
And you can get development environments that will target multiple devices, but some applications, especially when you’re looking at security, really you have to do coding for the device itself. You would write different code for the iPhone or the Android environment. Maybe different. Sometimes you’re referencing hardware capabilities. So the general mobile app, yeah, maybe there’s development software out there that you can build one that’s kind of hybrid and will straddle most of the target environments. HTML5 makes that much more possible than it ever was before. But you also get this situation where some apps actually can’t do that; that means that you’re actually doing the same work multiple times for each device that you’re targeting, and that’s not going to stop people claiming that they have a right to bring their own device; it’s not going to make any difference to that, so you can’t really get around it.
Apparently, analysis of mobile apps indicates that they drive more sales, right? And this is a strange kind of [inaudible] the website and the mobile app as, if you like, complement. The apps drive more sales. The websites are better at picking up new customers. Apps are better at retaining customers that you’ve already picked up. Customers spend an awful lot more on websites than they do on apps, but customers spend more often on apps. And that’s a really weird thing, and that speaks to the fact that if you’re going to build something, then you probably need a website incarnation and a mobile app incarnation, if you’re expecting it to be widely used. And that is, in one way or another, that’s kind of a dramatic expense to add to a software project, which in any case might be doing quite a lot of other stuff.
As a general idea, a website is a catalog and an app is a loyalty machine. Mobile app developments – and this is just to kind of break the problem down – different development environment, different problems in terms of hardware, different user interface design principles, and ability, you’re going to have to have offline capability – ’cause a lot of apps people expect to be able to use them if they’re disconnected – they don’t want to lose the data; some of the data’s got to be stored locally. You’re building a different app than you might build, let’s say for the desktop. And then, you’ve got the mobile back-end issue, there’s going to need to be middleware there, there’s going to be security procedures there. Quite likely there’s going to be a service-oriented architecture in the background, where you’re knitting together various things. And what this is saying is you don’t just take some team that’s used to developing applications on the server and stuff. Throwing them a mobile, you really need mobile developers. And people with mobile experience.
Anyway, having said that, just one more thing to say – above all mobile apps are, in most cases, a customer touchpoint, so they’ve got to be really good, because a customer will judge the company on the basis of the mobile experience, or it’ll be affecting their judgement. And in some cases, as I mentioned, the mobile app is actually what’s determined business success; it can be the thing that really makes an organization. And of course, it can be a damp squib as well.
And having said that, I shall pass the ball back to Eric.
Eric Kavanagh: Fine, and I’ll hand it right over to Bill. Bill, if you want to go to Quick Start there and share your screen?
Bill Ellis: Yeah. Here?
Eric Kavanagh: That top left-hand corner.
Bill Ellis: Yeah. Thank you for the instructions, I appreciate it. Robin, I really liked your discussion, it was funny. I’ve been, worked on a virtual team for 18 years now, so I think I can count myself as part of the mobile workforce. Sometimes I worry that I’m going to see, if I have an after-work function, I often have to get dressed to go to it. (Laughs) And I maybe start to lose perspective on what “dressed” is, so anyhow. (Laughs) With that, let’s go ahead and get started. I want to confirm that maybe Eric could just chime in and tell me, you can see my screen OK?
Eric Kavanagh: Yeah, looks good.
Bill Ellis: All right. So, my name is Bill Ellis, I work with IDERA on the Precise product line, and we’ll talk about enabling mobility. And we’re really talking about measuring it, and making sure that it’s working to your satisfaction. One of the big points there, was that it’s something that people kind of interact with, with your company. In a way, it’s very intimate – the phone is right in somebody’s hand and so the impression, the speed, makes a big impression on all of the users.
So, this was a customer experience I thought I would share. They had a go live, it did not go well. And because the initial load test did not fully reveal changes to the underlying application infrastructure, and so, one of the things that I like to stress is with mobile, whether application or HTML5, there’s also a lot of technology that that is dependent upon. Starting with the network, into the web server, into business logic, into messaging, and if they’re making a purchase, you know, a significant business transaction, they’re interacting with the system of record.
And kind of ironically, when we were getting started we encountered a couple of network issues, so all of this stuff is very pertinent even to delivering this webinar, itself. And so, you could have one application, at least six technologies, numerous end users, and just answering even the simplest questions is very difficult. Is an end user having a problem? What is the problem with an application stack, what code is causing the problem? And so, kind of getting a handle on those things is indeed not trivial.
Now, what we’re going to do, is we’re going to take a look at some measurements that were taken at a site, to help discern where problems are within the application stack. And what we’re looking at here is a graph, where the Y axis is response time, the X axis is time across the day. And the stack bar graph is a measurement of where the end-user transactions are spending their time. And so you kind of get a nice trend here, and then it’s going up and up and up. And it basically is the demarcation of the cutover, and so, consulting the stack bar graph, you can start to see that there’s a lot of problems in the J2EE tier. You’re also seeing issues in the web server tier, and then there’s some pretty big lifts, actually in the database tier as well.
And so, now that we have identified that there’s multiple tiers, with multiple problems, we need to go a little bit further in to find out exactly what’s going on, in order to have an intelligent response to this new usage pattern and this very slow, we’re talking about four or five X slower performance. And so one of the first things we want to do is say, “This is one transaction,” and so we’ve looked at the scope on the left-hand side of all of the transactions and they can, consulting, it’s really easy to look at the response time bar graph to basically see that you see in that same client web server Java for certain transactions more than others, database time. But it’s really across the board in terms of all transactions.
And this is looking at users, and so you start to get, this is a global deployment, so you’re looking at the primary continents within the world, so it’s all users, all locations. This is a global problem, it’s happening, so that starts to isolate, it’s not one or a particular group of users – it’s something that’s more happening on the data center side. And so we start to diagnose, well, where in the data? What application tiers? And so we start to look at average response time is building up, also layered over that with the number of executions, to kind of get an idea about scaling. This is very interesting – the lower half actually shows the longer-term history, and you can see very high access counts, but the other side of that is the number of concurrent connections is relatively low. After we switched over to a mobile HTML5 application, the number of connections more than doubles at a much smaller – we’re talking orders of magnitudes – it’s 100 times less accesses, so we’re not scaling; we’ve got at least double the number of connections to what we had had previously. So, we’re starting to discern what are the new demands that the mobile application is putting on the underlying infrastructures.
So, let’s go in even further, ’cause we need to isolate where issues are occurring. And so, here, you’re basically looking at kind of things scoping up, and we really didn’t need this bar graph here to say that we’re not meeting our SLAs, but we can easily see that in the upper graph. But we’ve got a secondary confirmation in terms of execution counts for SLA non-compliance. Now, here, we’ll actually start looking at locking, and this is within – this happens to be WebLogic but within the business logic tier. And you can see here, and this might be a little bit hard to read, but you’re pushing on 31,000 lock acquisitions for an aggregated lock time of 12 hours, 30 minutes. So, this is clearly a huge problem.
Now, the lock impact shows us that there’s always some derivation of the 80/20 rule. It’s really down to one method, one group of methods that’s really causing the issues. Now we’re starting to isolate problems within a particular tier. So, we’ll go in a little bit further, and here is the messaging system. And we start to see this, the over time graph that I’m circling in the top left, you can see the rough response time is going up, and the pink, the key, this is actually showing queuing and there’s actually a very different queuing that’s happening, that’s being pushed up, due to the number of connections. And so the messaging system is doing a lot more work; there’s a lot more – if you make an analogy to that grocery, there’s a lot more carts in each lane at the checkout counter – and that’s what’s pushing the queuing up, and you can see that most clearly in the domain. Each of the domains is seeing very, very high queuing.
So far, I’ve identified locking within WebLogic, I’ve identified queuing within the messaging system, and this happens to be Tuxedo. And then, what we’re looking at here is a similar type of analysis, but we’re looking at execution states within the system of record. And this happens to be execution states within Oracle. The reason why we focus on time is that time has two excellent properties. Number one: it is the way the end users and applications experience performance. Number two is it measures resource consumption. And so it will automatically identify where the bottlenecks are. And so I can see here, at the database tier, that I’ve got additional I/O time, so I’m stressing the storage sub-system. Every tier is the dependent upon the downstream tier, so the database is dependent upon storage. I can also see that within the database time, I’m doing locking. So, I need to get a little bit more granular before that information becomes a little bit more actionable. And so, let’s go in, peel back the onion yet another layer.
Now, this is actually a look at execution counts, the Y axis in this count, this is in thousands, you’re looking at 9,000, nine million, and so the execution count is also going up and up and up. So, the new mobility application is stressing the application a whole bunch of ways. Locking, just to recap: locking at the web tier, queuing in the messaging system, additional execution count on the database tier, additional I/O, additional locking within the database tier. So, we’re, I’m actually impacting every tier within the application spec. And so, it’s very important to be able to have metrics from every tier within the application stack. Here, I’m actually subdividing the database activity into program, and I can see that I’ve really got two programs: the turquoise color maps the application lock. And so, this one, the distribution server as application lock, the app, this is the mobile part, this also has application lock. And you can see there a number of these are bottleneck on storage itself.
Now I’m getting, peeling back the onion to see what I can do at every single tier. And the reason I’m doing this is a lot of people look at this from a capacity planning standpoint. And most of the cloud services, they talk about expanding servers, CPU and memory. The other side of the coin is just as important, is to application code that’s executing and driving the consumption of those resources. And when you know about the application code, you now can address capacity by processing efficiency. So, you have both sides of the same coin, and it gives IT professionals additional options for solving the problem. It’s not just add more servers, it’s also what can we do to clean things up and operate more efficient? The old “Work smarter, not harder.”
So here, we can actually, Oracle has a neat thing called Modules and Actions, where you can actually start to document the code, and so you can also get into looking at things another way, like here, the application lock that we saw? Well, that came in through the expense sheet code, it also came in through the distribution server, and so those are the two primary drivers of that new locking. And the new storage is coming through the online system, and so you’re starting to really build a profile, where the drivers are for this additional resource consumption. It’s another thing to be able to pinpoint the drivers in the underlying code. And so going into this, I think we looked at this expense sheet, and so we go in here.
Now, looking at the underlying objects that are being exercised, you start seeing this message log. Well, every time they do messaging – and we saw that it’s going up by a multiple – we’re actually touching this message log table and you’re going to actually see in a minute that that’s actually causing a lot of the locking within the database tier. So these new usage patterns are having a big impact up and down the application stack. Now, over on the right-hand side is the SQL code, and so this is actually the application code and we are tracking what SQL statements are doing by execution state. And so, it’s very easy through the color coding to see which SQL statements are involved in those locks. The reason that this is really vital is that if you go to your DBA, and you say, “Hey, we think there’s a problem at the database level.” They might just look at the database and it might look pretty the much the way it ran yesterday.
But being able to correlate the way the application is using the database, then, they can pinpoint the exact SQL statements that they should focus on, and then they can get into some of those advanced practices, looking at execution plans and all of those things that they can tweak, to make the system of record run much more quickly. And so, the correlating doubts of the code, it’s really vital to enable the technology experts to be able to solve and remediate underlying problems. Now, here, we also talked about storage – here, you see the number of physical reads, you can see when that happened, and this starts to get into the hardware architecture, because when you’re planning out evolving a system, one of the things you might choose to do is you can choose different types of storage, and they have a very different expense profile. And in certain cases it’s going to make a good sense to upgrade and pay for flash storage; if I’m doing a lot more random reads, then that flash storage is really going to pay off for me.
And so, the overarching message of this is that with a new application puts new demands on the system, and the underlying application stack needs to evolve to address those needs. And you also want to look at what those needs are and can the code be tweaked to make it be more efficient? And finally, down into the CPU, you can see on the cutover period, we had been running at approximately 10 percent and then, once with the new code, we’re at 4X, now we’re at 40 percent, and this is really important for physical as well as virtualized environments to make sure that you have adequate server resources to fulfill the needs of the application. And so, here’s just a more of a close up, so you can see some of those numbers little bit upfront. Interesting at the server level, the memory consumption hadn’t changed that much, but certainly the number of CPU cycles demanded had.
And this basically just is a recap of looking at the expense report, looking at the scaling, the fact that the number of executions actually went down, but the execution time went up. And so that showed that under the mobility, the expense component of the application was really having trouble. And that’s definitely going to have a user effect on things, ’cause if you can’t do your job, people are going to basically just stop using the mobility. And the nice thing about mobility is it really does empower the workforce productivity, and that’s very good for paychecks and so forth, so you definitely want that to roll. Now, we’re looking at the same thing here, just from a location standpoint, so that’s Europe and the Middle East, Asia VPN connections and then headquarters itself. And the United States overall. So, we believe that one way to get that valuable information at every tier of the application stack is through the precise product line.
I’m going to just very quickly, Robin and Eric, I’m just quickly kind of just provide an overview of what Precise does, and why it’s designed the way it’s designed. And what happens if end user is trying to do something, there’s a lot of technology in the data center, the end user really does not care about, they just want do their job. Meanwhile, you have a lot of people in IT, well intentioned, very smart, but they aren’t even aware of a problem until this end user reports, if they report. And then, a lot of times this is going to kick off a very expensive time-consuming ultimately frustrating process, where people look at a subset of the application stack, but it’s very difficult to answer those basic questions about who, what, when, where, why.
So, what we believe is by measuring the end user transactions starting in their device, through the network, into the web server, into the Java, capturing that information we can answer the questions of who, what, when, where, why, provide recommendations, but probably the most important thing is completing the feedback loop. We all need feedback to improve, it’s the only way that you know that something’s going wrong. By having the history put into a centralized repository, it provides one sheet of music for everybody to read from. And so, it becomes very easy to figure out where problems are, so once again, the design is about measuring the end user transaction; this is going to identify slow transactions, segmenting it, this is going to tell what technology is a problem and then providing an expert view on each of the individual tiers so that you can figure out what’s happening. Precise is going to provide a learning as well as reporting and dashboards for all stakeholders, whether you want to have just an overview, or if you want to have a deep technology view of what’s going on.
Now, what can happen, kind of like a day in the life, either you as an IT specialist could call an end user, or sometimes an end user could call you. Log into Precise, you can focus once again, the Y axis is response, the X axis is time across the day. Here, we’re each sub-state, so you have client time, web server time, Java, Tuxedo, database time. Down here you have the driving transactions, you can bring up a menu to identify a particular end user, and this way IT has the ability to address that particular end users’ issues. And so you can see exactly when they were busy, you could see that they were using content management you can focus in on that transaction and then Precise is going to give you an analysis of that transaction.
The percentage at the end is added by percent, Precise, and that tells you how much time, but a percentage of time, spent on that individual step, down to individual SQL statements, this is the context. And one of the things that we say is that everybody has tools, but few shops have context. And context enables the Java administrator to focus on the application code, the DBA to identify like in this case the particular SQL statement. And so, with that information it gives them a lot more visibility on how to address the underlying root cause for the particular transaction that was impacting the particular user. So, you really laser focused on the root cause. And you can analyze SQL statement, where did it spend its time, well, executing? And just by contrast, a lot of tools like Enterprise Manager just to pick on them. They’re big, they can take it. They look at things from an instance perspective, and that’s not enough focus really to get into these applications.
Typically, your OLTP mobility applications are going to be low latency, high throughput, so focusing on the top ten list, that’s a start but it’s really not good enough for this type of application. And then, the other thing is that particularly for internally hosted applications, identifying by user ID is really vital, because it’s not just about the application and the infrastructure, it’s also about how the end users use the application. And the end users typically have much better behavior when you’re able to identify them. And so this is just kind of a screen of different transactions and the client experience, and then sub-segmented, (laughs) I guess I’ve been speaking for a little long. Little tired out here; I’m going to plow ahead.
Here, we’re looking at a dashboard that we put together that would show alerts and then show different tiers of the application stack. Here are your web servers and you can verify by response time execution count that things are load balanced. You can look at browser accesses, you can look at keep usage and garbage collections, make sure you have that nice saw-tooth pattern, that you don’t have a memory leak, etc. And the idea of this is to provide a little bit of a more technical dashboard of each of the components within the application stack. So, the Precise product line offered by IDERA offers production monitoring, 24 by 7, very detailed information. It’s pretty easy to deploy this; you do not need to map transactions, whatever the end users do, Precise automatically connects the dots across the application stack.
If a downstream tier is not instrumented, Precise will recognize that and provide the in and out time and recommend that you instrument the downstream tier. And so, it’s very easy kind of time to value; we are very strong on the database, this is IDERA’s kind of claim to fame. And the reason it’s so important is that every significant business transaction interacts with the system of record, so the database becomes the foundational performance. And so the other tools on the marketplace, they do an OK job, but OK isn’t really good enough; you really need to know exactly what’s happening with the SQL statements. And we do a lot of advanced things, that are too much for this, like keep a SQL statement history and track execution plans over time. And so, that’s an area that we can explore further, if you might be interested.
So, with that, that’s the Precise application performance platform, we invite you to request an additional meeting through the idera.com website, if you have additional interest in the solution and the topics that we discussed today.
And, Eric, with that, I think we’re still under the wire, I’m going to pass the baton back to you and Robin. Thank you.
Eric Kavanagh: No, that’s fantastic and I love the content that you’ve put together here, because you do a fantastic job of displaying how complex the environment is underneath the hood. And of course, the whole job of Precise, the purpose of Precise is to help navigate that complexity and understand what’s actually happening and be able to take some actions to improve something. And I’m just kind of bewildered at how complex it is. I’m guessing that Precise also allows you to identify certain patterns of behavior and then name them, or at least record them or bookmark them or something like that, is that right?
Bill Ellis: Yeah, one of the things that’s going to happen, is you don’t want to go chasing your tail; you don’t want to just go spend a lot of time on a one-off. So, you would want to look at what are the patterns, what are the trends, because there’s a lot of technology to manage. And so one of the things is to prioritize and be able to rank, know where to spend your time, know what needs to be honed. And you also want to take a conservative approach of lower risk and lower cost. You don’t want to necessarily make an expensive global change, without having assessing or having a very good sense of knowing that it, this will indeed help the issue. So, know what’s happening over time and this trending is vital to intelligently addressing the underlying issues.
Eric Kavanagh: That makes complete sense. And how big a deal is virtualization for being able to see what’s happening, and then, are you coming into organizations that are using containers – using Docker for example? And how would that impact what Precise is able to do?
Bill Ellis: Yeah, so the word “container” can mean different things according to different vendors. And so, we work with VM, almost everybody uses VMware – I consider it the de facto standard at this point; I know that there’s competitors out there. And we are expanding what we support, but VMware is the dominant, within the Oracle stack. There’s containerized databases and so all of that is very important to be able to evolve your system very quickly. It’s also really important to know in a virtualized environment when the physical host is not able to meet the needs of all of the guests’ containers, because each of them are competing for resources.
And one of the things that actually happened internally I was surprised at, is that we had actually within IDERA so many idle VMs, but each of those idle VMs consumes resources, that they started to cause an issue overall for the VMs that were actually being used that were important to us, conducting our business. And so that was kind of an interesting thing. Now, we don’t support every technology under the Sun; there is a support matrix associated with this solution, and so that’s one of the things that we would want to drill down on, for a particular prospect or particular customer, just to make sure that we can meet the technology needs and the individual technologies that their application stack is running under.
Eric Kavanagh: Yeah, that makes a lot of sense. From your experience what are some of the major forces now that are driving challenges on mobile? When you and I talked before this webcast a couple of months ago, you made a really good point about how just the functionality and the layout of an iPhone or some mobile device can be a real challenge for the business, because suddenly the end user can’t figure out how to get some specific process in the workflow done, right? And so, to that point, what you’re enabling in mobile app development is you’re showing the developers where the problems occur and then you can map that back to what the app is doing on this particular device, or that particular device. And that’s very useful, right, for the developer, because now they can see what is causing the problem, they can make some change to the app, to solve that, right?
Bill Ellis: Yeah, it’s kind of the overlaying of incredibly high expectations – everybody expects everything to in a sense just work, but there’s so much variety out there. You have all of these different smartphones, they have different screen dimensions, and then, you have different vendors of communication, the Verizons, the AT&Ts, the Sprints, those are just the popular ones in the United States. And there’s just so much variety out there, it’s like well, how do you wrap your arms around all of this, to kind of start to discern where the issues are? And so, there’s a lot of metrics that are available and one of the things that our product management team has done, is attempt to pull in the metrics that are most important or most needed by the IT team, to be able to make intelligent decisions.
And so, it’s kind of a challenge and we do our product is like the marketplace evolving and so we get feedback from our customers and there’s always enhancement requests, so “Hey, this additional metric would be super helpful to us.” So, our product is evolving just like the marketplace, but if I had to say, actually Eric, it’s really interesting to me, is that whole expectations thing. People are like, it used to be back in the day people would wait five, seven seconds for a screen to pane up, now it’s like one or two seconds, people are like “Oh, this application doesn’t work at all!” (Laughs)
Eric Kavanagh: That’s funny. It’s so true!
Bill Ellis: It’s crazy.
Eric Kavanagh: Yeah, it’s a bit unrealistic, honestly. And I think that maybe we’ll start seeing a bit more realism around that topic, but nonetheless it is a fact of life that people have very, very high expectations. And I guess, Robin, I’ll bring you back in real quick in the last couple of minutes here. I loved your assessment of website as a catalog and app as a loyalty machine. And to that point, what we’ve been talking about here is how to enable the developers of these apps to understand what’s happening: Is it usable? Is it not usable? And what can you change to adjust that? And to Bill’s point here, just a second ago, the cycle time on fixing that problem has really shortened, right? It’s just not like it used to be – you’ve got to fix that quick. Or you’re just going to have a huge drop-off in use, right?
Robin Bloor: Yeah, there are a whole bunch of other things that are played into this, so you’ve got this agile development and you’ve got expectations in a lot of places now, that you’re going to release a new version of something that’s in the process of being developed, or in the process of being changed, every couple of weeks. And that puts, it makes when you think about it, if you think about the deployment environments, and you think about how big the stack is when you’re getting into mobile, you’ve actually got multiple potential devices on the end node, and then you’re going to have middleware in the middle. And you may well have [sower?] underneath and underneath the [sower?] you may well have databases. So, you may be touching many, many applications; you may be touching multiple databases and you may be doing very complex things in terms of security. And it’s all got to work, and the expectation is it’s going to work reasonably well.
And the amazing thing is sometimes it does, but my thought about this, is if you really, if you’re building mobile apps that are really key to the success of the company and a lot of them turn out, a lot of these things really are. If you’re doing mobile maintenance on oil rigs and oil pipelines and things like that, it’s kind of got to work. The consequences of it not working are just kind of dire. And if you don’t have this ability to actually slice up the application and know where things are going wrong, ’cause most of it is performance. We’ve got really good test harnesses nowadays, so yes, there are bugs and bugs do get through. But mostly if something’s going wrong it’s performance issue. And if you can’t put the stethoscope in 18 different places, then it’s really difficult to pin down what’s going wrong. And you also have the network’s a factor in this, and you also have the reality that any given component in an application can be stressed at different times of day, because of the nature of that particular application. You’ve got to have sophisticated monitoring tools if you’re going to stand a chance with all of that.
Eric Kavanagh: Yes, I would have to agree and I think that’s really the strength of Precise by IDERA, these days. And Bill, I guess just any closing comments from you? I think that this technology’s fantastic. I also realize that as a user of this technology, you really need to understand the complexity of information systems and the dependencies and be able to figure out where, when and how you synthesize all this information to assess what’s actually happening. And that requires an intelligent and trained human being, and frankly, it’s one reason why I’m not at all concerned about machine learning taking away jobs. I think machine learning could be very useful underneath a technology like this, to identify common patterns and then make suggestions to the end user as to what might be happening here. But what are some closing thoughts from you about really getting the enterprise the importance of having this kind of troubleshooting capability and what should they know about that, besides what you said already?
Bill Ellis: Yeah, so Eric, I would agree with you there’s a tremendous amount of complexity. I believe the Precise product line by focusing on the metric time, that a user who can read a stack bar graph can use Precise successfully and I just want to say thank you to the participants and to you and Robin for hosting today’s webinar.
Eric Kavanagh: You bet! And like I said, we will host this archive for some time now, so feel free to share it with your friends and colleagues; we do archive all these webcasts. I sent a link to the slides a few minutes ago, feel free to check that out, but great job again, Bill, today. You really know your stuff; it’s always fun to work with a professional like yourself. And I think this is really going to be the enabling technologies for the mobile workforce! So, thanks for your time, folks, we’ll catch up to you next time, take care. Bye bye.