The internet is out of (your) control
Applications at a crawl may not be down to anything you’ve done wrong, we learn.
Show Notes for Series 02 Episode 04
Every time you connect to the internet (you’re probably connected in several ways simultaneously right now), are you aware of the intricacies of what’s actually happening in terms of data moving across the worldwide networks? Most people aren’t, and that’s where Thousand Eyes comes in. It’s a company whose technology can tell you exactly what’s happening, where, and creates (literally) a picture of internet health.
There’s more to it, of course, than pretty illustrations of traffic flows. For instance, if the app you’ve lovingly crafted or website you’ve spent months designing starts running slowly, or even grinds to a standstill, how do you pinpoint the problem? There’s so much outside of our control, many problems exist “out there” and, in many cases, there’s nothing we can do about it. In fact, we can’t even pinpoint where the issue might lie.
At least, we couldn’t until very clever people at Thousand Eyes gave us that ability. Their technical chops were recognized a while back by a certain Cisco Systems, who completed their acquisition of Thousand Eyes at the beginning of August 2020, in a deal rumored to be worth in the region of one beeellion dollars.
With knowledge comes power, and armed with the knowledge of where your organization might be providing a third-rate customer experience, you can do something positive about it: change to a more local CDN, invest in your streaming technologies, or start replicating services to better connected hubs. So, before you fire the CTO, listen to this episode of the Tech Means Business podcast, where Yogi Chandiramani (VP Solutions Engineering, EMEA at Thousand Eyes) will let you know what the practicalities are of today’s internet.
The internet connectivity report Yogi mentioned:
Yogi Chandiramani at LinkedIn:
Joe Green of this parish on LinkedIn:
Full transcript available.
Joe Green (host): Hello there, welcome to the Tech Means Business podcast. This is a series of podcasts that explore the technology that underpins today’s businesses and organizations of whatever size. In each episode, I talked to a subject matter expert to their field who’s got something to say on that particular area: insights, expertise, and experience to share.
Today we’re looking at the subject of the internet. You can’t cover that in 24 to 32 minutes, I hear you cry! But wait! What we’re talking about is how the internet and the stuff “out there,” if you like, not your local network, can affect all the apps and services you use and offer. How can we stop issues beyond our control, having an impact on what we do? To help us explore this area, I’m joined today by Yogi Chandiramani of ThousandEyes.
ThousandEyes is a company whose technology is capable of mapping its way through the internet, plotting a course through the myriad routes, routers, cables, bottlenecks, and superhighways. And then with that route map, if you like, in hand, making sure your organization’s not affected by the issues that are out there on the internet becomes a whole lot easier. Now, ThousandEyes has clearly shaken a few cages recently. It’s been acquired, in fact, recently (that deal’s gone through this very month) by Cisco for an enormous pile of cash. So I was lucky enough to be able to speak to Yogi before he drove off in his Lamborghini. I’m kidding. Of course, although I do hope he got something from the deal!
I started by getting straight to the point with Yogi and asked him if an app seems to be performing badly, it besmirches the good name of the business, and it’s irrelevant, really, where the issue might be. In fact, the app itself might be firing well on all cylinders, and everything to do with the app also running perfectly: the internal networks, the connection to the cloud, databases, clusters, microservices, routers, whatever APIs, can all be working absolutely superbly. But if there’s a bottleneck or some kind of hold up, miles away completely outside your control, it still gives you and your application and your company a pretty bad reputation.
So how can a company like ThousandEyes get the message over that, that it, in fact, can help solve those issues?
Yogi Chaniramani (guest): ThousandEyes was built to empower customers to see the internet like it was their own environment. And as the internet is vast and composed of thousands of service providers that can interconnect and any of which can impact the experience of the users. The ThousandEyes platform collects all this different information from various vantage points. And similar to how you use your GPS to avoid traffic, ThousandEyes provides visibility into outages and bottlenecks. And that’s really what we are helping our clients to gain access to, which is giving them in the sense a Google Maps of the internet or weather map of the internet, so that can best decide where they should be connecting: from which cloud provider that should be selected in different regions.
There is an assumption quite broadly that the internet is the same everywhere you connect from. The reality is that the internet is different depending on your location, depending on the service provider you select, depending even on the time of the day, and depending even on the day of the year. And being able to understand all these different factors, how they could impact the user experience really enables organizations to use a GPS, which is a much smarter way than just looking at a static map of the internet.
How networks are connected and being able to connect through: this was even more important, I think during COVID-19, you know, under the new remote business realities that we all have been experiencing. We’ve actually been more and more dependent on the internet to deliver the mission-critical apps; everybody has been working from home. And it’s not only been business people working from home, but it’s also been students working from home and sharing the same router, the same router to connect to the internet. I’ve been working from home, and I’ve been sharing the same infrastructure as my teenage kids. So we’re connecting to the E-learning systems. And obviously, you’re fighting for bandwidth.
We were fighting for connectivity, as we all wanted to get real-time applications running, such as video conferencing and other types of apps. And I think what we really learned and what we at ThousandEyes, as we observed learned during this time, is that there were really big shifts of traffic that were happening overnight, while people started to work from home. They were relying much more on the local ISP and the local infrastructure. And as a result, where you would normally see traffic going from the corporate networks towards the cloud application, it shifted fundamentally going from the whole networks to the cloud applications. And as a result, put actually a lot of pressure, if you will, on to the infrastructure.
The internet did survive, you know, thankfully. But there were pockets of outages that occurred because of these massive shifts of traffic, which really explains or rather, which really tells us that it’s more a systemic change that can happen.
So if one service provider is being challenged, it would challenge other ones, as the connectivity would be backed up into other service providers. So that was absolutely interesting to observe during the last months, and how those shifts really impacted the traffic on the internet.
Joe Green (host): It is, of course, fiercely complex, isn’t it? That you’re talking about backbones, ISP routing tables… I mean, for me at least I think it’d be very tempting to overshare if you like, that technology information. Isn’t it, you know, kind of a better idea (or more difficult, I guess in some ways) just to show a picture and say, here are the hotspots.
Yogi Chandiramani (guest): The internet, you know, is really a technology or rather it’s very technical, right? There are so many things that happen when you’re connecting from a location to another one. Our job at ThousandEyes is to provide context and to provide intelligence into what it really means so customers can really make actions. Our job is to simplify the complexity that comes through all these different interconnections, OSI models, as you mentioned, and so on. But really, say you know the problem is actually happening at this side of the infrastructure, and it is impacting these and these types of applications. As a result, you may want to engage a third-party to help resolve this problem, as we live in an environment which is highly connected, not only in terms of connecting users to applications, and application to application but also highly connected in terms of service providers and different vendors we interconnect with.
So we rely on our local ISP, the local ISP relies on their own transit speeds, you know, the wholesale internet, if you will, side of the business. And then we rely on cloud providers and so on.
And also, typically, clients will outsource the operations to another vendor, and how do you ensure that everybody is seeing information exactly the same way, so that they can work to focus on solving the problem as it has been seen, than to not having all the data to truly look at it. So it’s more than just giving a map or just giving a picture or just giving a hotspot onto into a world map if you will, like the weather map; but it’s more about being given, what is the context around this outage, and why is it happening so we can solve it more fundamentally.
But also, most importantly, it is to involve the right people too, to solve it because we rely so much on the internet and on those cloud applications, that makes it even more critical to be able to effectively work together and make it really efficient for everybody to leverage the internet and the applications in the cloud.
Joe Green (host): Okay, so let’s say I’m running a startup business, and I’m offering it as a SaaS, Software as a Service, you know, the kind of thing that people pay for monthly say. But gradually, the metrics are getting worse that the application, or whatever it is I’m offering, appears to be slowing down — just as an example. And you know, if it’s retail, you might say something like you’re getting high levels of cart abandonment if you like. So how do I go about identifying an issue that isn’t inside my network, inside my control, and then start to actually ameliorate or address that particular issue?
Yogi Chandiramani: Indeed, and you know, what, what really comes is, first of all, is the problem within my network, or is the problem outside of my network? But at the end of the day, it is my problem I need to solve because it is impacting my business, and it is impacting my customers, my users, employees, and potentially as well partners.
And this is where you know, ThousandEyes can really help organizations gain those insights.
The first question that our customers ask us that, or rather, they want an answer for, is the problem with my environment? Or is the problem with the network or the application? Being able to give that answer first of all really helps to identify who we should be working with to solve this problem. And that is point number one but also giving them information about where the impact is happening. And it’s happening at this specific router, which has got this IP address. And this is where we are seeing outages occurring from a packet drop and higher latency and different types of metrics.
But identify precisely which equipment or which service is impacting the overall experience. Do this as being able to see and get the same visibility into the connectivity between the users and the apps, even if the connectivity is going on outside your control, outside your network, by being able to get that same level of insight. So you can actually ask the service provider if it is a network issue, to solve the problem and to be able to look into it.
And to make things a bit more complex as well, what we often see is that the issues are intermittent. They’re not like completely broken, or the application is completely broken or is working very, very well. It’s like, sometimes it’s working, sometimes it’s not working. And that becomes very, very complicated to understand and to resolve. And this is where we can see the different patterns with insight can really help organizations to pinpoint precisely what they need to optimize.
And you know, they need to bring more capacity because there’s one of the service providers that is having some capacity, some contention in the infrastructure, or if it is happening at the application level, because, I’m getting much more traffic and maybe I need to optimize my application infrastructure altogether.
Joe Green (host): Okay, so to swing this round to ThousandEyes itself. If once a problem has been identified, say, does ThousandEyes manage to remain neutral? Or do you suggest a partner company, you know, a new cloud provider or a CDN, something like that? Or you remain kind-of “Switzerland-like,” neutral and giving objective opinions, and just letting the enterprise make its own decisions?
Yogi Chandiramani (guest): We will provide insights to customers, and we provide them the data and the metric. At the end, it is going to be for customers to make the decision. If they want to make a change, or if they want to adapt the architecture, we will provide recommendations in order for them to follow and to leverage our experience and how we’ve seen this done. But indeed, it’s going to be more like, you know, providing this intelligence to our clients and then from there being able to leverage it. And in order to optimize their infrastructure as required. It would seem it is right because it’s important to align as well with their business goals and their business strategies and what they’re looking at, as well as doing it longer term, sometimes, can be a bit more complex when you’re working on a big digital transformation initiative and wanting to really implement specific initiatives or specific projects.
Joe Green (host): So I’m guessing then that bringing in a company like ThousandEyes can stop a good deal of the kind of finger-pointing, blame gaming, that goes on! I mean, if, for instance, I don’t know, our database team has a particular toolset, it can point out that there’s a problem somewhere else, it’s not their fault. But that kind of subjective pinpointing of an issue, I guess, can be dismissed as, you know, part of the whole blame-game. And it’s that objectivity, valuable to end war-rooming and escalation and that kind of thing.
Yogi Chandiramani (guest): Exactly. And I’ll share with you a story which happened to me last week. I was working with a client that was deploying a business-critical application for them in China. And China is a big growth market for them, and they were seeing performance issues with this application. And that in no way it was if it was the network if it was the application, which was poorly designed or needed some optimization. And they were really scratching the head about trying to understand what they should be looking at. And they implemented ThousandEyes in China. And they were able to see precisely what the problem that we’re having is, that it was more the way they’re accessing the CDN application, which was connected to the main application because today, applications are complex. It’s just not one side; it is multiple calls, which are happening through APIs or other types of services altogether and bringing everything together and packaging everything together into the browser.
But one of the applications which were pulling some information from the CDN was connecting outside of China, and obviously that was not mainland China. Obviously, that was not working very, very well. As we got in, we helped them get those insights.
We were able to tell them, look, the problem seems to be here as your application is being served, the content is being served, outside of mainland China. And that’s what is impacting the performance. Within hours, they were able to go back to that provider and say, it seems you’ve got a configuration issue, which they did. And it got it fixed. To your point. You know, if you didn’t have that insight, how long would it take to actually identify those types of issues? It would have taken war rooms; it would have taken escalation to FDA, taken a lot of frustration most probably, about where it is we’re not talking the same language, but being able to correlate all the different layers altogether and be able to see one single view. The problem is there. And that’s what you need to look at. That’s a context which turns into action.
Joe Green (host): So I’m intrigued. Who are the people who pick up the phone at 3 am to shout at people, shout at people like ThousandEyes, I guess, you know? Are they the systems administrators, deep in the bowels of the data center, are they the suits and ties, the application owners, the operations directors, the VPS, and the CEOs? Who’s actually on the end of the phone to ThousandEyes and companies like you.
Yogi Chandiramani (guest): It’s really two folds. Um, you know, the first one that we would typically get a call from is going to be from the operations team. It’s going to be the IT operations or network operations that are facing a challenge, and they just don’t know how to solve it. And they’re coming for help in terms of getting in those insights and being able to better understand what they need to do. And those are indeed, you know, the typically the teams that get blamed when a problem happens, right? They get finger pointed at, hey, please, you need to help me solve this problem! But also we do get a lot of CTOs, heads of infrastructure, and even digital officers.
They do understand the importance of being able to predict the digital experience, being able to take action before even there has never been a problem or even potentially an outage, and be able to really design, if you will, the infrastructure, and the operations in an optimum way. And as a result, we head into the process and work with the head of infrastructure, the Digital Officers, or even architects in order to integrate that this visibility is built-in into the project from the very first day.
Joe Green (host): Now is part of the challenge one just of language? If, for instance, you’re taking a situation that’s terribly complex and translating it or parsing it into “normal English” I mean? This is a situation that I find a great deal as, as technology permeates more or less every single area of life, never mind of business.
There are layers and layers of abstraction, if you like, layers and layers of making it easier to address technology, it becomes more and more complicated to, kind of, peel away those layers and actually explain what’s going on. I mean, is that part of the challenges as you see it?
Yogi Chandiramani (guest): Yeah, it is. It is very, very much. The fact and it is also even turning this into business KPIs. In order to achieve those business KPIs, what does it mean from a technical standpoint to make it happen?
If an organization wants to have a digital service all the time, operating at specific levels, what does it really mean? And how do you ensure that it’s going to happen? This is really where it comes in.
You know, more and more, in terms of discussions and being able to detail precisely about how the processes can be optimized and can really help the business get a competitive advantage. Really, your technical problems are going to happen; operational issues are going to happen. But the question is not if it is going to happen, but what you’re going to do with it, and how you’re actually going to solve it and how you’re going to react.
So you can be ahead of the problem, and you can stay resilient at all costs towards the service that has been provided to customers, to employees, and also to partners. And these are really the conversations that come in, come in more around resiliency, optimizing digital experience, and also optimizing about how customers will leverage those tools as a team.
Everything today in the world has become an app, and we use so much of those and become really a norm, so that experience becomes absolutely critical.
Joe Green (host): Is there a role do you think, Yogi, for teaching the basics of network infrastructure? I mean, it sounds terribly dry. I know. But take, I don’t know, my kids, for instance. You know, they’re very much shielded from network realities, by — well, I would call them abstraction layers. They probably think of them as, as things like, well, it’s just a point and click, it’s a GUI, you know, it’s a simple thing to do. But actually, what’s going on under the hood is incredibly more complex and complicated. And so is there a role, do you think for some kind of teaching of that kind of nitty-gritty of, of networks and network infrastructure?
Yogi Chandiramani (guest): Definitely, you know, I think everybody believes they know how the internet works in and out, you know, my teenagers as well, you know, they think the internet is just simply a network you get connected into, and that’s it. But the reality to your point is that it is so complex, and it is so sophisticated. And with the fact that we rely so much on the internet for everything we do, it’s absolutely fundamental to everybody to understand the complexities that go with it.
And yeah, just wanted to circle back something I mentioned a bit earlier is that the internet is not the same everywhere; it’s not the same at home.
Indeed, you know when you are going to be in your garden versus close to your access point and so on, because that the experience is going to vary. But also it’s going to be different in various parts of the regions; of the world. And also, it’s going to be different depending on the cloud provider; you’re going to select differently based on the ISP that is going to be selected as well. And being able to understand why that is the case is, I think, absolutely fundamental. Because everything relies we rely on this infrastructure, we rely on this network, which is key. And sometimes we just take it for granted. And taking something for granted is fine, but understanding why it is so, so we understand the dependency we have towards it. And the reason why it is so important to us is absolutely fundamental. And also, it gives us a better understanding of how to leverage it. If we understand it more deeply about how it really works!
It gives an option for our future generations, and even ourselves to better leverage the benefits that it brings and you know, how it works and how we can really use it at a much more optimal way and deal also with some of the issues with latencies and so on that could occur. So, for me, it is absolutely fundamental to understand it, at least at a broad level, how the internet works.
Joe Green (host): Now this year, ThousandEyes has been being acquired by Cisco Systems. And in fact, I think at the beginning of August, the last ‘I’ was dotted and ‘T’ crossed of that particular deal. What does that mean for ThousandEyes as a company, and what does it mean for your customers? Are you going to be subsumed into the morass of Cisco, or will ThousandEyes exist in five years’ time as a recognizable entity?
Yogi Chandiramani (guest): The feedback that we’ve been getting from our customers and partners is that this announcement is very, very exciting. By bringing, you know, the power and the strength and networking and application performance of Cisco’s portfolio, and also all the experience and all the partner ecosystems that they’ve been building across all these different years and alongside with the insights and the visibility that ThousandEyes provides.
The feedback has been that you know, it is very, very exciting for our customers in terms of seeing new use cases and being able to leverage the technologies and much broader ways than they had thought about it before. The feedback as well from partners is very, very exciting because they’d be wanting to embed a different type of you know, a different type of offering and packages, etc. I think it’s good to make things much more easy on that front. But indeed, the many many things which are being worked on as we are discussing today, and more will be shared in the coming weeks and coming months.
Joe Green (host): So Yogi, in pre-COVID days, it would be at this point that I’d give my guests the opportunity to promote or plug some event or you know, a slim volume of poetry! Now clearly, there’s nothing like that going on IRL as the kids say, in real life at the moment, but how is ThousandEyes at the moment getting its message over? Is there anything where you can, for instance, point our audience to go and get some more information?
Yogi Chandiramani (guest): Yeah, definitely, definitely we’d love to talk about it. So during the lockdown, when the lockdown rather started, what we decided to do at ThousandEyes is to provide a weekly report of how the internet is holding up. It’s called the internet report. And this is done through a blog and also a video on YouTube. And it’s really, really interesting to see how the internet evolves week after week and what are the different learnings we get out of, out of all the different analyses, and the insights we’ve been able to get.
I would definitely encourage our listeners, our audience, you know, to go and look into the internet report. And that gives you much more context on helping to better understand, you know, what is really happening on week by week. So that is some definitely learning which I think is key to get to.
And of course, we do have multiple virtual events like everything is becoming more and more virtual, but where we do have demos which have been shared, there’s one which is coming up next week, and which we schedule every two weeks every four weeks around different topics as well. And the goal is to provide insight into one topic, but also share it with a real use case and with a demo and with an example of how that could be implemented. So this could also help to better understand how this technology works into action and how it could help different use cases.
Joe Green (host): So as is all too often the case, I’m afraid as the music comes up, that’s the sign that time’s run away with us. As usual, we’ve got far too much more to talk about than we’ve got time for, unfortunately. So it only really remains for me to say Yogi Chandiramani of ThousandEyes. Thank you very much for joining us today. And I hope that you, the listener, can join us next time on the next episode of the Tech Means Business podcast.
11 November 2020
28 October 2020
13 October 2020