Benefits of an IoT Optimized Cellular Network (with Felix Hsieh and Richard Halliday)

This is a recording of an internal conversation between Soracom product manager and customer reliability engineer Felix (Shay) Hsieh, and Soracom partnership manager Richard Halliday during one of Felix’s monthly AMA office hours calls in which he invites anybody in the company to ask any questions they have.

In this week's episode, you’ll get some very specific examples of what happens when teams of cellular network engineers partner up with cloud engineers to break new ground in the world of connectivity operations.

Some of the best IoT product experiences come down to anticipating the small things. That means building the tools and networking options that IoT engineers, hardware developers, field support technicians, QA specialists, or IT teams can use to do their jobs faster, gain more control, or streamline non-negotiable things like security. For many deployments, that means addressing old problems in new ways.
This can include:

• Managing the time, expense, and security risks of static IPs
• Navigating the number of hoops you need to jump through to configure VPNs and private networking over cellular
• Diagnosing network traffic issues and perform packet analysis over cellular networks
• And more.

I’ve listened to this conversation half a dozen times and always seem to take away something new about the differences between traditional cellular data networks and examples of what the future looks like for IoT-optimized cellular networks and how engineering teams are modernizing the experience for those responsible for building and maintaining networks of IoT devices.

Transcript

0:00

Welcome to conversations and connectivity. I'm Ryan Carlson, your host. This is a podcast for IOT professionals and the IOT curious, responsible for growing executing, or educating others about conductivity and the myriad of moving parts. Associated with conductivity operations. This recording is an internal conversation between Soracom product manager and customer reliability engineer, Felix Hsieh and Soracom partnership manager, Richard Halliday during one of Felix's monthly AMA office hours calls in which he invites anybody in the company to ask any questions they have in this conversation, you'll get some concrete examples of what happens when teams of cellular network engineers team up with cloud engineers to break new ground on the world of connectivity, operations. Now some of the best IOT product experiences come down to anticipate in the small things. That means building the tools and networking options that IOT engineers, hardware, developers, field support technicians, QA specialists, and it teams can use to do their jobs faster, gain more control or streamline non-negotiable things like security, which means addressing some of the old problems in new ways that can reduce time, expense and risks of things like static IPS, doing something about the insane number of hoops that you need to jump through just to configure a VPN or a private network through cellular providers and even consider basic tasks that you'd assume that would be easy to do like diagnosing network traffic issues or performing packet analysis. Nope. Not over cellular networks. These things can be tricky. Now I've listened to this conversation at least half a dozen times and on how has, seems to have a new takeaway for me. There's these differences between traditional cellular data networks and examples of what the future looks like for IOT optimized cellular networks and how engineering teams are modernizing the experience for those responsible for building and maintaining networks of IOT devices. And now. Onto the conversation between Felix and Richard.

Richard: 1:54

Hey, Felix, how are you? How's it going? Yeah. It's all right. I did have a couple of questions for you I'm trying to prioritize the things that we feel like are the most valuable to the customer. So if you had to think about three services that you would say yep. These are by far the most valuable, to customers. Off the top of your head, do you have an idea of what those would be?

Felix: 2:20

Are you speaking in terms of what we offer on our platform?

Richard: 2:23

Yeah, for sure.

Felix: 2:25

It's always different for every customer because every use case has something that's a little bit different, but what I think is universal is, the cellular connectivity, lifecycle management tools that are just built into our console. These are things like being able to activate deactivate and do all that type of controls. and the biggest part that I think is really, useful for a lot of customers is the level of automation that comes with that. So you can go to a lot of other, competitors and, they will give you many of the same lifecycle controls. Some of these competitors, they don't give you quite as much flexibility as we do. but one of the key differentiators I feel is that we allow you to, automate a lot of this stuff. You might think it's trivial. If you only have 10 Sims to just log into your user console and select a SIM and then deactivate it when you don't need it. But if you are trying to do that with a fleet of 10,000 or a hundred thousand Sims, suddenly you have to pay someone full-time salary to just sit around and do that. And that's where TCO topics come into the conversation. And, we can get, someone on the phone, and help them set up those automation, features. Basically the first one I would say is the lifecycle together with the automation. In north America there are a lot of use cases where customers end up finding Soracom Napter remote access to be quite useful. and we do have a little bit of a challenge in terms of the positioning because most customers are coming from the angle where well, at the very basic level, Napter is providing remote access. So you've got your device deployed remotely somewhere and you need to connect to it as opposed to having the device pre-programmed to connect back to, like to phone home or something. and so if something goes wrong with the device and you don't want to fly someone out, you want to be able to remotely access it and see what's going on. that's what Napter does. And the challenge we have with the positioning is simply that a lot of people with these types of use cases where they need that remote access, they're familiar with having a static IP and there are a lot of security issues. It's also very costly. So that's why we created Napter remote access now there are a couple of limitations and that's where there's a little bit of an art to make sure that we really do Focus on the main situations you would need a remote access, as opposed to some of these outliers that Napter doesn't really support. But, I think there are a lot of different verticals and a lot of different scales that actually do find Napter quite, compelling.

Richard: 5:01

One of the things that we're working on is this, business continuity product for high bandwidth usage and one of the applications that we're thinking about is really around, retail fail over. Sure And so one of the requirements often is this notion of having a static public IP address. To what extent can we offer Napter as an alternative what are the challenges we're likely to run into in that kind of application

Felix: 5:27

do you think? fail over is a fairly plain or a fairly unsurprising customer requirement in terms of what a customer might be building, but the exact ways in which they are expecting fail-over to work can have different implications on whether they should be using Napter or if they should be using a difference Soracom service such as a VPG. Napter really intended for, the remote access part, of connecting to a device. Depending on situation. Some customers might, have, if they have, for example, a, some kind of a network device that is providing internet connectivity to, sub devices, and they want that sort of redundancy with the fail over, over cellular. it really depends on, what the connection from that network device is actually connecting to. So if their main internet connection is just providing internet connection, a Soracom SIM is also just providing, an internet connection. So fundamentally there's no real difference. The kind of main points where you might run into is if your main internet connection is used to connect to a VPN server. Then when you fail over to the Soracom cellular connection, that is going to be a different IP address. And so your VPN server is going to be like, Hey, this is a different IP address. I don't recognize it. So I'm going to block the connection and that defeats the purpose of a fail over. So in that situation, rather than using Napter, we would recommend, our virtual private gateway and then we do have the, the, fixed, global IP address option for that. but there are a lot of different flavors of fail-over. So whether VPG is right or not, and definitely VPG also does have some additional costs, beyond, the individual SIM itself. it's not necessarily a one size fits all. It's everything and the kitchen sink. that doesn't mean that it's for everyone.

Richard: 7:18

yeah. Okay. cause again, I'm just trying to think of that's one of the potential blockers we saw because some other companies do offer that as an option. And I understand. Yeah. And I understand that we've made the decision that the security risks are too high. because it's one of the first things we're going to get asked, we have a partner who, while they describe themselves as a national distributor of IOT hardware, they also, do a level of service and integration. And one of their biggest customers is a gaming customer. They do a ton of, lottery machines in north America. you can imagine, The lottery jackpot gets above a hundred million and all of a sudden. But every lottery machine and every Vigo down for, I can't re I never remember the statistical but for every second they are unable to process a lottery ticket. It costs them a ridiculous sum of money it's like, you know, like, oh, on average, people lose$20,000 worth of revenue in a year because it fell apart. Not lottery machines it's like a whole other order of magnitude, fail over, has to be as close to instantaneous as possible. And, and I pretty sure that's an application where, it's going over a secure connection, A private network, Makes sense, but typically at a future date, but I it's good to have that at my fingertips so that we can talk a little bit about when it comes to those partner conversations.

Felix: 8:43

We can dive deeper there. Yeah, I'll provide, just a kind of high level way of looking at this. we, between VPG and Napter remote access, we have two extremes, on the Napter and the, The, the static IP address, we do end up actually providing, a static IP address, but it's just set up in a way that it's a, it's an ephemeral IP address. It's not actually permanently fixed. We only lease it for, Well, we set it up for eight hours for now. If we need to set it up for a longer weekend, talk about, potentially, increasing that we said eight hours, just because, in most of the use cases that we're usually talking about, we don't expect people to be remotely connecting in, to, to devices for more than eight hours. but the reason that it's on one end of the extreme is that it is a. a single IP address for a single device. And. It's pushed further along the extreme because that single IP address can only exist for eight hours at a time on the other end of the spectrum, we have the VPG virtual private gateway and the high level way of thinking about this as well. Some customers, maybe this might be relevant for the gaming company that is managing all these lottery machines where they say, okay, we do have 50,000, lottery machines, spread around the country. We do need a static IP address, but fundamentally we don't really care about having a dedicated IP address for every single machine. We just need a dedicated IP address because maybe we have a VPN connection or something like that. So we can authorize that connections coming in from this IP address. Are, I'm allowed to connect to our VPN. So what the VPG does is it takes all of these devices and it puts them underneath the same IP address, the single static. now that's where if you do have use cases where it's I only have three devices, you don't really want to be paying the whole cost of a VPG just to have. Three devices share that IP address. If you have a thousand or 10,000, that definitely makes a lot more sense. we do have a couple, finer points that essays would normally talk about such as reliability and scalability, but, that's the high level now the middle ground, where most people come into the conversation is one static IP address per device fixed permanently. And so we have the, these suit sort of two polar opposites and nothing in between. and depending on the use case, we, it might be easy or to nudge one way or the other.

Richard: 11:17

Yeah. that's how I appreciate you framing it that way for me, because it, that makes the distinction between those two, that translates much better for me. and I'll be able to speak to it a little bit more intelligent in that way. So that's helpful. and so those are two, a, your top three. Yeah. what's number three on your all time favorites list of, features you think customers are most likely to be attracted?

Felix: 11:41

Number the spot for number three is pretty, pretty hard. yeah. I definitely have three or four that I would put at the bottom of the list. the one, there is one that I think, So Soracom Peek It's not a total hit for every single customer, but for the customers that run into these types of issues, it is like a mind-blowing sort of experience. when you're like an IOT developer, and you're doing some at home tests and you're just doing rough prototyping, and maybe you're even at this stage where you haven't implemented cellular connectivity and you just have wifi at home or Bluetooth at home and so you're like, okay, that's good enough for now. I'm only just doing this testing And when you're using a local network, such as a wifi or Bluetooth connection, there are a lot of really great tools that you can use to inspect how your device is behaving. and a lot of these, center around packet capturing. Packet capturing is basically nearly ubiquitous all over IT. In the context of local networks or corporate networks. So if you control your network, then you have that control to be able to sniff packets, to capture packets, inspect packets, same thing. If you're in an office, your it team has the, Control of the corporate it network and therefore they can, roll out these tools for capturing packets and inspecting and seeing if there's any malware or anything like that, happening in the corporate office, Where this breaks down is when you switch to cellular, you no longer control that network, at least in the traditional sense, because it's AT&T that controls the network or Verizon that controls the network. And if you wanted to go to AT&T or Verizon and be like, Hey, so my devices are having some problems, they're deployed all over the place and I'm having trouble figuring out what's wrong with it. I'd love to be able to just run some packet capture on your network AT&T is going to be like, I don't think so. We don't want you mucking around our network. if you can even get hold of them. Yeah. you'll be waiting for three, four weeks until we got a response. and so that's where Soracom Peek really shines and people are like, wow, I had no idea that this was even possible. it just blows my mind that I have the ability to capture packets on a network that I don't own, that I don't control. And of course this is just based off of all of the technology that we have. It's this is where, if customers that don't run into any issues, when they're launching into cellular or scaling up, they have no reason to use peak, but, When customers do run into problems, it's a really powerful tool. I wouldn't say that it's necessarily the, in that sense, like the tool that Al that would attract a lot of customers, but, again, being able to demonstrate that, Hey, this is a network that you don't control. However, you still have a tool to be able to do deep network analysis. Like packet capturing which peak is, for a number of engineers and it professionals, they will look at that and be like, that is different than what everyone else is able to provide. And that's in some ways, even easier to explain the power of that, or the implications of that than trying to explain the implications of VPG, you know, security and private networking and that type of stuff. So, Peek is not very far reaching, but I think it, it, It's a really good example. of the type of, control we give back to customers over their cellular networks.

Richard: 15:22

Yeah. And you said that the customer is going to be primarily interested in that is the one that's doing their own development work on they're probably starting off on a network, that isn't corporately run. Or centrally run?

Felix: 15:35

Yeah. That's why I hesitate to put Soracom Peek at number three, because not everyone is coming from that exact stage. the people who are DIY buyers who are hobbyists, who are also have some sort of business, expertise in order to turn their weekend project or their sort of POC into an actual viable business model. that is not everyone that does not fit every single profile. And so the ability to communicate that for those people to be able to look at Soracom can be like, Wow that's something I need to get into my toolkit. that's not really going to message to that many people. If I were to say, okay, let's extend the list to number four and five. I probably would add in something like Soracom Harvest together with Soracom Lagoon and then maybe Soracom Beam, but then similarly, they have their sort of limited audiences. And depending on, what exactly the customers are dealing with, or the innovators are dealing with at this particular time. so around number three, positions, three to five it's it really gets a lot more contextual,.

16:44

Hey, this is Ryan butting in here as this is an internal conversation. There's a little bit of lingo here. There is beam, which is a protocol conversion tool. So it allows devices to use unencrypted traffic like HTTP or TLS. And it adds the encryption as it sent over to cloud services. So that's way of increasing battery life, reducing power consumption and reducing the data that's sent over cellular. And if data equals dollars, man, it means you're spending less dollars. There is then harvest, which is connection, validation, and simple data visualizations. It's a way of getting to do proofs of concept. Of taking device data, stores it into its own little S3 bucket, and then allows you to like, look at the data without having to set everything else up on the other end. So. Great for early testing. And then there's lagoon, which is full dashboarding service. Now this is not the core services of what soar com does. But it's already there. It allows for people to set a venting, alerting, sharing out different graphs. Great for getting internal support or socializing an idea. And then, you build out what you need to, when you get to the actual full production of whatever the IOT project is, kind of an r and d team's best friend i'm going to turn you back over to richard and felix

Richard: 18:02

Yeah. so it sounds to me like Peek would be a good service for us to spend time promoting at a resale partner like, Digi-Key where a lot of those folks are coming to Digi-Key because they're doing prototype work or they're doing development work. It's an early stage. Many of them will do, as you say, a personal project or a bench project in their garage, even though their full-time job may also have to do with technology and development and that sort of thing. So that might be the kind of situation where we spend some time talking about why this is a important and be different than any of the other connectivity providers

Felix: 18:46

Yeah, definitely. I think, especially in the context of Digi-Key, who are the people that are likely to be browsing through the catalog in the first place, those are going to be people who have, technical product managers or technical project managers, as well as supply chain, procurement, folk who have sort of an eye towards bomb costs and trying to minimize the bottom line something like Soracom Peek Does actually connect towards, lowering BOM costs or lowering TCO, mainly because, if you're like, wow, I can choose this SIM instead of that SIM and if my engineer has ever run into any problems, they can just turn Peek on and start capturing packets and inspecting what's going on. And then a feature that none of these other Sims have, let me go check and see like maybe the SIM is a little bit more expensive, but let me go check and see with my engineers, if that would be something that they would be interested in.

Richard: 19:44

Okay. All right. That's helpful. you talked about, harvest and lagoon. what's the difference in the top five, but the use case, the value to the customer is different because

Felix: 19:58

right. when I say the indicative type of person that would find peak, valuable is the sort of like garage DIY or we can hobbyist, it's not exclusively to them. they're actually a larger teams. If you have a, an engineering team that is responsible for developing device firmware, and, Uploading these firmware to all these devices, they're updating all the device from where the devices are already deployed all over the place. If they do mess up in their firmware, then they can use the packet capture to inspect that and see that. So it's a little bit more broad reaching in terms of where Peek can deliver value. it's just that, I think it's a little easier to conceptualize it as the DIY the weekend hobbyists who doesn't even know that, in general cellular networking, you don't have control over cellular networks. because they're on their home network, they expect that they can control their networks. And what they don't know is that when they make that leap to cellular networks, If they don't choose Soracom, they're going to lose all of that control overnight. what I really like to think about harvest and lagoon is it's not at an end product that the customer really wants. It's more of a stepping stone product. harvest and lagoon are number one. There were good, services for number one, capturing data and the number two building visualizations, using that data. But the real reason that, customers would use harvest and the lagoon is because. It exists and if it exists and the pricing is reasonable, then there's no reason that they have to go and build it themselves. If it exists. And the pricing was ridiculously high, then of course, a lot of customers would be like, maybe it's just better for me to build it myself. So that's where we're playing around with this, R and D cost discussion. So the way that I like to frame it is that if you have, a small startup, hodgepodge, team of, various engineers, That, have this awesome idea for an IOT product, and they've decided to make the commitment to leave their full-time job in order to focus on this startup of full time. Those folks are running on a limited budget. They still have to pay their rent. They still have to pay for food and utilities. they're all digging in and there's a limit to how much, how long they can do that. this is what we would normally call in the VC world. you're runway. so depending on, how fast you're burning through your cash and how much funding you have that defines how many months you have to get something up and running. when you look at it in that term, it's yeah, you can definitely go and spend three or four weeks to build your own data capture and your own dashboard. Or you can just turn on harvest and lagoon and use that three or four extra weeks to do something else more productive. And so we can actually help them increase their runway. And in, in some situations, actually, if you have a really early stage customer, that's saying I have this awesome idea. I don't have all of the engineers that I think I need. You might have a conversation where they think, I know that I need to capture the data and I know I need to create some kind of a dashboard. That's what I have in my mind, where I have all these devices and some kind of dashboard showing me all their information. And we can say, Hey, hang on a second. Before you called and hire an engineer. Why don't you pay 5 cents a day for harvest? that's a lot easier. And then you can save that salary to extend your runway a bit. so that's where harvest and lagoon have a wildly different. value now, harvest and lagoon are great, but I think a lot of customers, once they get to a 5,000 SIM scale or 10,000 SIM scale and above the probably want to graduate from harvest and lagoon. And at that scale actually roll their own solution. harvest and lagoon is great for maybe tens of thousands of pieces of data at a time. But when you're dealing with millions and tens of millions of pieces of data, there's this great partner that knows that, their software is amazing for giant data sets and then we can pass them through over.

Richard: 24:09

So I, it's hard for me to be technical, but it's very easy for me to visual. Have we ever constructed something that shows the life cycle of a company and basically mapped our services to the lifecycle of the company? that could be potentially very visually powerful.

Felix: 24:24

Absolutely. I'm trying to use, visual assets or visual communication to get the same points across. the general idea that we had with this concept was just think about a factory where you have an assembly line and the assembly line is zigzagging all around the place. And at the very beginning, you have your raw materials at the very end, you have your finished product the product in this case is the IOT. Product that the innovators are creating. The assembly line is their IOT journey. And they're getting from just an idea phase all the way down to production and scaling up. And what we're doing is basically taking a pit stop at all these major points throughout the assembly line process. so we have some raw materials at the beginning, starting from the design. And then, communicating some of the places where we can deliver value when you are at the design stage, then you get to the prototype stage. then at the launch stage then at the scale stage and

Richard: 25:24

then exactly. Yeah. So, so I like this idea, right? Because it, because it, it follows a journey that, uh, a customer at any stage is going to recognize.

Felix: 25:35

Right. Exactly. And, and the, the, I mean, there's a little bit of a. The trick to this is trying to find the common denominators that really hit home regardless of industry, regardless of vertical, regardless of application. We would definitely want to come in here and give some like, uh, uh, clear examples, especially in the prototype phase. This is where we would definitely want to name, harvest and lagoon, uh, and say like, don't worry about hiring additional people before you even make any commitments to developing your entire cloud architecture. We have tools like harvest and lagoon that can help you get up and running. Literally within seconds, you just log into the console, turn it on, and you've already got that entire functionality, at your disposal. versus for example, you're going to the scale stage and that's where we start to talk more about the automation, and the fact that by automating your SIM life cycle, your device life cycle, you can reduce tCO, you don't have to hire people to constantly activate and deactivate Sims, things like that.

Richard: 26:41

Yeah, that's great. and it doesn't matter where you are in that company life cycle. We've gotten something for you.

Felix: 26:46

I'd like to add one additional viewpoint for your consideration. it's not by any means anything that you should incorporate. it's not something that we, carry front and center in the rest of our messaging. But, the way that I like to think about another way that I like to think about it, if you are an IOT developer or an IOT innovator, IOT, isn't something like you come up with an idea, you produce a thousand units in China, you put it in Amazon, you sell it and make a profit and you move on. It's not that easy. IOT is hard. If you are in IOT, you're in it for the long run. and if you're, if you don't make it, you're going to be, running the risk of not being successful. So you are committed. What you need is someone else who is equally committed. This diagram demonstrates Soracom's, commitment to being in the long haul together with you. And you don't need to use all of them. And we're not going to push any of these that you don't need. But the point is, as you're going through your journey, it's going to be a long journey. It's going to be rough. We're right there with you. Even our CTO will jump in on sales, customer calls and work together with you. So that's, that's a little bit of a twist in the messaging that I think, can be very helpful, to demonstrate. You're looking at Digi-Key as a catalog of different, part numbers. can you understand that when you select the Soracom SIM, as opposed to a different SIM, that you're getting people who are equally committed to a long-term game as you are?

Richard: 28:22

Yeah. and that's exactly why I love the way that you're thinking about that, because that's exactly why this kind of storytelling is so important because. What has started to happen is while you certainly get some supply chain professionals who are in there punching in a part number and then looking for similar part numbers and trying to drive costs a lot of attractive. Is people doing research because they can get multiple suppliers in a single location. But at the end of the day, it doesn't end with a SIM card. It only starts there. You've only just started. Exactly. I have a gong way to go and we know exactly.

Felix: 28:57

That's yeah. Precisely.

Richard: 29:00

Nice. thank you for going through that with me. And this is going to be super helpful So that's great.

Felix: 29:04

Thanks for hopping on and asking all your questions.

Ryan C: 29:06

There you have it. Thank you so much for tuning in. And now this recording has been brought to you by Soracom a global conductivity service provider that believes that the fastest way to cost savings and scale is when customers are in full control of their connectivity operations. Experience, your own self-service pay as you go global connectivity without a contract today at Soracom.io and believe it or not signing up for an operator account takes less than a minute. Once you got a SIM in hand, you can start testing. It's going to work on all of the networks anywhere you go. Anywhere in Canada, anywhere in the United States, 260 other countries with over 5 million IOT connections out there and growing.

29:43

Soracom is. One of just a handful of. The IOT optimized cellular networks that did way more than just the standard mobile virtual network operator. Which just adding like billing or consolidating, usage onto a single SIM, but it goes well beyond that, by taking core cellular services and giving you the same access that. The big names have when they get a look at their own network to perform that packet analysis, to turn things on and off to quickly switch things to. See exactly where something is connecting on real time. All of the same tools that the big network operators have you can have because. Well, again, the core infrastructure now lives on AWS, which means you can go from one SIM to a million Sims. Without any delays? No excuses, no extra servers. No wait times, no long contracts. None of that. If you've enjoyed this podcast. Please go ahead and leave us a review on Spotify or apple or wherever it is that you're listening to this. We'd really appreciate that. Thank you so much.