Many organizations are struggling to get a hold of managing cloud costs, and do not always have insights into their usage until the billing arrives. However, it is not the most efficient way to manage the resources that are being consumed when you are only analyzing costs and usage after the billing period.
Amberflo aims to change that, offering a cloud-metering, usage-based pricing and billing platform to help organizations take complete control of their cloud costs. In this episode, Swapnil Bhartiya sits down with Puneet Gupta, CEO and Co-Founder of Amberflo, to discuss the challenges organizations are facing with managing cloud costs and how Amberflo’s platform can be used to help solve these problems.
- Many companies are moving to usage-based business models but it can be difficult to keep a tab on usage. This is where Amberflo’s cloud-metering, usage-based pricing and billing platform comes in. Gupta introduces us to the company and tells us what it is aiming to achieve.
- Gupta takes us through why the usage-based business model was introduced and how the nature of cloud computing is elastic so resources also should be elastic allowing for them to be dialed up or down as necessary. He explains to what extent costs are a major concern.
- AWS is managing its own internal cloud costs, and according to Gupta, gets ahead of the bills by using cloud metering. He goes into detail about how that helps teams measure what they are consuming without having to wait for a bill. He explains how this is used to contain your cloud costs.
- Gupta believes that cloud metering has both a cultural aspect and a technological one that are intertwined. He feels that there need to be best practices for how resources are consumed and monitored but also for appropriate technology tools. Gupta explains how these need to work together in order for organizations to stay ahead with cloud cost management.
- There are a variety of tools that help with cloud management costs, but many focus mainly on analyzing the billing rather than managing costs before the billing. Gupta discusses what sets Amberflo apart from other tools, going into detail about the metering capabilities and the pricing and billing application.
- Gupta discusses what the dominant business model will be, not just for cloud but also for software in general. He believes it will be usage based not only because it is fair and transparent but also because the technology foundational stack is going in that direction too. He goes into detail about the benefits of usage-based costing for customers, giving examples.
- Cloud resources are being spun up and down on a second-by-second basis, so having the ability to set thresholds and view the resources being consumed and manage them appropriately is a game changer. Gupta discusses how the metering tool is helping organizations get a hold of their resource consumption.
- Gupta shares his tips for how organizations can embrace the usage-based model and move from the traditional legacy model to this new model. He explains the importance of putting in place the foundation of a metering system and starting to get the raw intelligence and insights.
The summary of the show is written by Emily Nicholls.
Here is the automated and unedited transcript of the recording. Please note that the transcript has not been edited or reviewed
Swapnil Bhartiya 0:00 Hi, This is your host Swapnil Bhartiya. And welcome to TFiR Let’s Talk. And today we have with us Puneet Gupta, CEO & Co-Founder, Amberflo. Puneet, it is great to have you on the show. Hi, guys.
Puneet Gupta 0:10 Great to be here with you Swapnil
Swapnil Bhartiya 0:12 Yeah. And so this is the first time I’m talking to you. And also you are the co founder of a company. I’m curious about the company itself. Tell me what problem you saw in the space that you wanted to solve, that led to the creation of this company?
Puneet Gupta 0:25 Yeah, sure. That’s, that’s our story. So what Amberflo does, what we provide is we enable businesses to charge and track usage. Okay. And what that is, basically, you see this trend now, more and more companies are shifting to what’s called a usage based business model. Right. So unlike a traditional subscription based business model, where companies charge you by number of users and a fixed charge per month, more and more companies are following a business model that Amazon Web Services pioneered, which is pay as you go, you only pay for what you use. So usage based. So we have a technology platform, we call it a cloud metering, usage based pricing and billing platform that enabled companies to charge and track one usage,
Swapnil Bhartiya 1:16 it seems that you know, not only usage, but the cost associated with cloud and cost is not just in dollar amount, but also developers time it is becoming a real concern. As you know, there is no, even an Open Source project called OpenCost, which helps with Kubernetes causes. And there are a lot of products and companies which are trying to do that. So I would understand from you is cost really becoming one of the biggest concerns when it comes to cloud adoption, because the thing is, unlike previous word, where if you need resources, you have to go to the team as them it will take five, six months, but you will get what you want. Here, you can get it what you want today, you can get it yesterday, and at some time you open provisioned, you don’t even use all those resources. So talk about, first of all, how serious is this concern? And why has it become a concern?
Puneet Gupta 2:04 Yeah, great question. So. So let me do it sort of bifurcated in two ways. So you know, because cost, and the pricing model of what you charge to your customers is kind of intertwined. But let me sort of draw a distinction. And then sort of perhaps help clarify, you know, what might be a good framework for, for customers or companies who are in the cloud to sort of think about a lot of this is, you know, just from my background, my co founders background, we spent few years at AWS, Amazon Web Services in the early days, I personally had a chance to with the team out there build a couple of tier one AWS services. So that’s sort of you know, where I got quote, unquote, religion about usage based pricing, billing, underlying technology primitives cost optimization. So we’ll cover all of that. But what you asked, see, let’s take a step back. If you were to ask that, why did even this usage based business model even came about? What was it even at AWS that they chose to launch with a usage based pricing model, right? Because think about it, they didn’t have to, they could have just as well launched an AWS S3 service and made it subscription basis, right. But they decided to do it on a usage base. And the underlying vector is that cloud is elastic by nature, that was the value proposition, right, that these resources could be consumed on an elastic basis, you could dial it up, or you could dial it down. That was the premise of cloud computing. That was the underlying thesis. And if that was the founding design principle of the service, that its elastic in nature, then naturally, you needed a business model, a pricing model that was also elastic, that would, you know, if you use more, you would pay more, if you use less, you would pay less. Now as things went on, clearly sort of cost became an issue because you could easily dial it up or dial it down, perhaps for lack of controls, people could you know, just consume more, and then perhaps, you know, leave some of those resources running. So costs started to become a discussion in cloud computing that, you know, how do I contain costs? So that’s one vector. Okay. And there are some tools out there that you could use to sort of get ahead and gauge a little bit, you know, what you’re consuming? So that’s one vector. The other is, okay, if you’re building in the cloud, so let’s say companies like snowflake, Twilio, so all these modern Cloud Native companies, how do you what is the business model that you are going to sell your services, products, services to your customers, on the backs of which kind of what kind of a business model and you will find that these companies who came in as we call them Cloud Native, and they build their entire stack natively on the underlying cloud providers, they also chose to sell their products and services on a usage based model, pay as you go, right? And what I’m highlighting is, again, that none of this is by AXA. And the reason why this happened, why Twilio, snowflake data breaks, and increasingly just about every other company is looking to launch their products and services on a usage based model is because the underlying usage, their own cloud resources are elastic basis. So to have a full alignment, operational efficiency from the front end, when I say front end, what your customers pay you, and what you pay in, turn back to the underlying cloud provider to achieve maximum alignment. You need to have an elastic business model to and throughout. So that is the place where we’re at today.
Swapnil Bhartiya 5:35 Excellent. Yeah. No, no thanks for once again, explaining that in detail. When it comes to cloud cost, some of these topics might seem off topic. But to be honest with you, I’m talking to so many guests, which is I mean, today, I recorded two interviews just about cost with two different companies. So which also seems that it’s becoming a really, you know, serious topic. So I’m trying to, you know, I might even make a montage of all the thoughts that are there. So I may ask a lot of questions there. One thing with the cloud, as that says things are so accessible, folks end up over provisioning things. Now, it’s very easy to over provision, but 99% time you may not even use it. There are tools that will help you monitor your usage, but they will not help you actually do anything about that. So talking about from you know, Amberflos perspective, how do you look at this problem? How do you help users to optimize their cloud usage?
Puneet Gupta 6:46 Fair question, very fair question. So, in fact, what I’m going to outline, let me back that up a little bit with some dataset, right. So let’s ask the question, who’s running the cloud really, you know, cloud operations at massive scale? And how are they managing their cloud costs? Would it be fair to say that, you know, I think AWS would be a good example. Obviously, AWS, you have internal teams who themselves are running large services. And they’re not just running blind, they must also be doing their own cloud cost management. So let’s ask the question of how AWS manages their own internal cloud costs? So first, I will tell you that this is not it is not an unsolved problem, it has been solved, and it can be solved. Okay. The artifacts are slightly different now to AWS as credit, and to their distinct advantage. You know, they started, they started with a clean slate, ground zero, right. So they have over the years build certain artifacts and tools that helped them stay ahead of this problem, say, okay, but I will tell you and your audience that whatever they have, it’s not rocket science is not exclusive to them. It is just Google business process, best methodology, and certainly a toolset that kind of works in cohesion, and helps them stay ahead. So let’s unpack this. Okay, so today, the cloud cost management tool that is available to most of us, is on the backs off, what many of these tools do they will parse your AWS bill. And they will tell you, you know, you know, here’s some resources that were not fully utilized. So you should dial them down. Right. But a lot of it is a lot of the tools out there today, based on parsing your AWS bill, right, or your cloud providers bill and asking you to take some actions based on what they derive from the bill. I will tell you, categorically. And factually, that is not how AWS does it. I can speak from firsthand information, as I mentioned to you, I lead a couple of services there. Cloud cost management is a big initiative inside AWS, all GMs managers have to own it, and have to continuously take cost out of the equation. So the way to do it Swapnil is you have to get ahead of the bill. Let me put it in more direct terms. If you are a leader inside AWS as a GM owning a service or whatever, you’re not waiting to get your AWS bill, to figure out your cost profile and how to manage that. You put artifacts and probes in your technology stack that gives you visibility on what you’re consuming ahead of the bill. And there’s a term for that. It’s called Cloud metering. So the teams deploy what’s called a metering service. And you basically have to usage instrument yourself what you are consuming don’t rely and wait for the cloud provide either to give you the bill, and then try to try and parse that bill because that’s totally too late. You have to put probes in. So you have to in your technology stack, you have to put a layer of usage instrumentation. And the word for that is called Cloud metering. Put that artifact in your technology stack, and quote, unquote, generate your own usage bill of your consumption before it even gets to the AWS bill. And once you have that dashboard, and that metrics, that’s how you contain your cloud costs.
Swapnil Bhartiya 10:32 Excellent. Yeah, as of course, I just said, No, you have the first experience being at AWS, yourself. So you do know that now, when you talk with folks, you know, should have the dashboard for metering, I also want to talk about you know, you always talk about build versus buy. But before I go there, I also want to understand, because this is also important topic is cultural versus technological technologies are there, but culturally, if teams are not prepared, before they are hit with the bills, so talk about what you’re seeing in that space, how much awareness is already there? How much do you see this as a technology problem? Versus a cultural problem?
Puneet Gupta 11:10 Yeah, another good question. So, the two are certainly intertwined, right. So there is a cultural aspect is, you know, you have to at least make the teams aware generate a level of awareness that we used to call that, you know, resources are ephemeral, they come and go, right. So there has to be a discipline, there has to be, quote, unquote, a set of best practices on how these resources are consumed, but then also how these resources are monitored, and then sort of kept track off, right. And, but see, culture and discipline alone will not cut it, unless you have a technology tool set that enables that culture, and that best practices. And then more importantly, that technology tool set is actually viable, it actually works. And it is delivered to the right teams at the right place. So they can make use of it. I cannot overstate the fact that you know, the reason why Swapnil we are still talking about cloud cost management and cloud, you know, AWS is 15 years or more in the working, you would think that, like I said, this is not rocket science, this problem has been solved. at AWS. No team, you know, is surprised inside AWS, you know, they are in control of it. Okay. So you understand what I’m saying, right? Because it’s been 1520 years in the working, you know, you don’t get a pass inside AWS oops, I didn’t know, you know, I hit a cloud cloud cost overrun, you are ahead of that process. So yes, it is a combination of business process, methodology, combination of culture, but that has to be backed by tools that, that teams feel empowered to use in the tool actually delivers. Okay, so I will go back to the right combination is, first give me a toolset that actually was really designed for this purpose. And metering is that artifact, right? You basically, wherever you’re going to consume your cloud resources, whether it’s an S3, put or get, whether it’s launching an EC2 instance, or storage or any cloud artifact, right? You have to put your own usage instrumentation probe. So you know that I’ve used this resource. And because cloud is all API enabled, nothing happens by itself, either it’s automated through a script, somebody has to take action to do something in the cloud, right? It could be a script on your behalf, but somebody has to execute that script, or some, you know, some API call has been made somewhere. So you can track it. And you can use it as an instrument, right? And then you have this discipline, where you have a dashboard. And on a weekly basis, people are monitoring that. And you get into a meeting on a week or weak bases, and you have a dashboard that you can rely for being accurate, that you can take decisions on the backs of that. That’s how you come full circle, and staying ahead with the cloud cost management.
Swapnil Bhartiya 14:05 Excellent, excellent. Once again, great equation. Now, since you’re talking about tools and instruments, let’s go back to the point of you know, there are a lot of Open Source projects, there are a lot of built in tools as well. And of course, there is Amberflo as well. So talk about the pros and cons of the benefit or, you know, building it damsels or using tools, you know, that you folks are working on. And if you can also talk about the edge that you provide, as compared to there are a lot of other players who are also offering similar, you know, solutions.
Puneet Gupta 14:40 very cool. Yeah, so two things. One is okay, respect to this Open Source, you know, where do we fit in? What’s the value proposition, you know, why you would go Open Source versus something like Amberflo. That part is, you know, fairly straightforward. See, Open Source tools are there but again, just like anything Open Source you may get a start. But ultimately, you know, a lot of this will just come down to management growth scale, and then kind of staying ahead of the roadmap. Right. So while I did say that, you know, we’ve been at this for now 15 years, almost nearing, you know, 20 years, when AWS first launched their service, I was back in 2006 2007. Right. So, but the problem has not really been solved. Because I think the approach today, unfortunately, for cloud cost management has all been on the backs of analyzing your bill. And somehow we’ve been stuck there. So it is a different paradigm. And there aren’t really a lot of tools out there. With the approach that I’m talking about, which is a metering approach. You could certainly build your own metering tool or metering system. And in that context, you could leverage some third party or Open Source tools to kind of get ahead. And guess what at some of the companies I mentioned earlier, like you know, Twilio data, bricks, all of these guys, they didn’t have a choice and they have actually built it. You know, Twilio, Jeff Lawson has also acted for AWS, so he had the awareness of what needed to be built to stay ahead of them. And they have actually built a metering tool in house. But it has gotten unwieldy, there’s so many people on the backs of her trying to maintain it. So you will eventually find yourself in that realm because it is not core to you. It is, you know, essentially an IT function so to speak. Right. And that’s kind of the value proposition for Amer flow. I mean, unless you yourself are looking to build a metering service and offer metering service to your customers, then metering is an IT project for you, right it is an enabler, it is a tool. So we are packaging that end to end, in a you know, self service pay as you go, full featured AWS styled, fully managed cloud metering service that we offer for the use cases that we just talked about that you can literally just enable. And then on top of that, we are also giving you a pricing and billing application. So now that you have a handle on what has been used by whom, when and where and how much, you can actually layer yourself a pricing model that you can project back to your customers, so they can consume your services AWS style in a usage manner. So that’s sort of our differentiation. So we are the only vendor who is offering you what we call a decoupled ala carte cloud metering service, you actually can use a metering service, independent of our pricing and billing service, just as I described, usage instrument, what you’re consuming on the back end on the front end, that’s cloud metering, its job is nothing but to give you a accurate system of record in perpetually in real time, what is being used by whom, when what were how much period, that’s the job of the metering service, do it consistently do it accurately. Now you’ve got that system of record in place, which is the metering service, you can layer a pricing plan on top and you can do other functions, and you can feed that data downstream into third party applications. That’s the value proposition of fabric.
Swapnil Bhartiya 18:02 As earlier you were talking about the pricing model, do you feel that going forward, the pricing model will change where you know, they will move to USPS. Where we will move to usage based model often recommended like for example, VMware is a good example. After this acquisition Broadcom is trying to you know, remove all the seat licensing and moving everybody to a subscription based model. people are panicking that hey, you know, we have already bought so many licenses. So this will be African recommend. But the point is that do you feel that? What is the ultimate pricing model not only just for cloud services in general SAS, but software in general. So that once again, the point is that you pay for what you use not just a bulk licenses, doesn’t matter. The license sitting there, nobody uses it, but you need it because that can be the service.
Puneet Gupta 18:55 Yeah. So Swapnil hands down, I cannot. Short answer. Yes. And let me so sometimes, you know, I get asked the question, okay, is everything gonna go to usage based? I think the right question is what will be the dominant business model? Right? And I can categorically look you straight in the eye. And I can tell you that the dominant business model is going to be usage based. Okay. But let’s challenge that assumption, right, easy for things to say for me to come out and say it. I certainly have my own biases. But here’s why. Okay, let’s first and foremost, I don’t think anybody can deny the fact that usage based is definitely far more fair and transparent. Right? The premise of usage based is what you pay only for what you use, isn’t it? Okay. Now, so that by just by definition is quote, unquote, sort of the Holy Grail, right? I mean, you and I can shake hands on that, that hey, yeah, that is fair. You know, when you if you’re offering me something, you’re saying, you know, I will only pay you as long as I use you, right? Yeah, that’s fair. Because if I don’t use you that I don’t pay. So that is the underlying premise of usage based pricing model. That’s why customers love it. That’s why things are going in that direction. The other direction we already talked about is also the technology, foundational stack is also kind of going in that direction, because companies are finding that, hey, my back end is elastic. So I have to ultimately align, you know, if some request comes in from the front end, that gets farmed out on the back end, I have to basically align the two models. So that’s also why companies are leaning towards usage base, it’s a win win. Okay. So there is no question that, you know, things are going to go in this direction, when there are multiple vectors that are going in there pushing things in direction when it’s like I said, Cloud Native, if you’re building on Cloud Native technologies, those are elastic, they will push in this direction. Second is there’s going to be a huge poll by customers, because usage based eliminates shelfware. Right, customers are starting to learn and appreciate Yes, nothing is more fair and transparent than paying only for what I use. And AWS and others sort of pioneered it. And it has now gotten into people’s psyche, consumers customers are gravitating towards that. Because if my business landscape change, if there’s a downturn, if there’s an upturn, if my company is going up or down, I should be able to dial down my contract value with you. Right, you cannot be holding me to a contract that was signed six months ago, let alone two years ago, three years ago. Okay. So for that reason, things are going to go in this direction, I will also give you a couple examples that we should learn more credibility to then so what we are seeing is not just desktop now in technology, I’ll give you a couple examples of a company called Metro mile automobile car insurance, okay? Charging you insurance by the mile. usage based, right, they put a little gadget in your car, and it tracks how many miles your car travels. And at the end of the month, they send you a bill based on how many miles a car travels. Companies like even as far flung and traditional domains like Caterpillar, John Deere industrial equipment, thinking about moving into a business model of unit of work performed by their machines, metric ton of dirt moved. Okay, it’s happening combination of IoT combination of the technology stack, giving users the flexibility, it’s actually opening up the total addressable market for them. By going to this model, it is going to be the dominant business model, The only hiccup is there not that many tools that companies who came early into this had to build it themselves. It is a fundamentally different model. This is not a step function. This is a fallout change. So as you call out earlier, there’s a cultural aspect to it, there’s a retooling technology aspect to it. Right. So that takes a little bit of time, but it is inevitable, it’s going to happen.
Swapnil Bhartiya 22:55 You’re so right. There are two things there. Number one is that of course, the the challenge that comes with the metering, usage based pricing is that you do need tools built in hook which can actually track the usage. And that is you know, a lot of onus falls on the those companies, they have to build those tools to be able to meet this second thing it will ask and I may be totally wrong there. It also becomes makes things more resource efficient, because sometimes you get it, it just sits there collects dust, you know, it’s not being used. But since you bought that license or subscription, that resource is allocated to you where someone else can could have used it, when you move to a pricing. If you are not doing something you will stop using it at that moment. Number one, number two is we are talking about cloud. But if we just look at on prem, we are on premise, not what I mean is and do you I’m not talking about the pricing model per se, but more awareness or state says to sensitivity about the sources that you are not using you should free them. So this metering can also be we can flip in a different way that if you’re not using a resource, you should release it from your bucket so others can use it. What are your thoughts about that
Puneet Gupta 24:10 person? And again, I cannot again overstate see metering is that artifact. So there’s a little bit of, you know, awareness that is now starting to emerge as more and more companies kind of follow the playbook on the usage based business model because they’re seeing that it ultimately delivers greater retention, higher net retention rate, all of those things. So people are exploring this, but metering has been sort of hidden in the background, right? Like you and I are talking. I would challenge you there isn’t one vendor out there other than Amberflo who’s offering you a ala carte cell service pay as you go, fully managed metering service in the cloud. pureplay AWS Styra So it’s new we’re sort of in the early innings as more and more companies are shifting like we discussed the companies who haven’t had to build it themselves. Okay, so, but metering is that artifact you You’re 100%, right? You know, cloud resources are getting spun up and spun down on a minute by minute second by second basis, that’s just the nature of the cloud, you need to instrument that, and you need to do it accurately. And then you need somebody, anybody inside the organization should be able to go into a dashboard, and see what has been used by, you know, who, when, what, where, how much. Now, again, within the realm of, you know, a security layer and access rights and controls layer, right. But if you have this platform in place, you have the tools. And now you can put a layer of best practices discipline culture, the tool can notify you, this is not being used by anybody, or I’m not taking any traffic. And you can set this is the job of the metering tool, you can set thresholds. If this goes above this, let me know if it goes below this, let me know. So that’s how you manage it. It is not rocket science, as I said, it is being done. You know, this is not left, which I it is not left to chance, categorically at AWS, I would venture to say a lot of the other companies will kind of build discipline around this,
Swapnil Bhartiya 26:05 as you’re saying, you know that ultimately, it’s about what will be the dominant pricing model. And you know, usage base is there. But we have a lot of industries, companies, which has been around for a lot longer time. And then we also have companies which are born in Cloud Native with era, which are much more versed with usage based pricing. So can you share a not necessarily a paid playbook, but some suggestions and tips on how folks can either embrace usage based model or how they can prepare themselves, not necessarily migrate, but you know, adapt themselves to move from the legacy model or traditional model to use a paid model? What are your tips there?
Puneet Gupta 26:45 Certainly, yeah, you know, there. So first is, you know, don’t be don’t be scared, don’t be paranoid. I mean, this is it, it can be achieved now, but there are some systematic steps that you have to take. And if you look around, there are companies who are doing it. And now they’re, I think enough case studies out there. But let me sort of give some, some highlight, at least to kind of get started on the right path. Break it into different chunks, don’t try to boil the ocean. The best advice I can leave you with is the starting point. So if you start down this path, this approach you are on the right path. And that is that when you are thinking about moving to usage based pricing, don’t start with your pricing model. First, don’t try to come up with what your pricing model will be customer facing. First put the foundation of metering, as we’ve discussed, put that usage instrumentation, technology primitive foundational building block, put that metering system in place, start to get the raw intelligence of how things are being used on your product, who’s using what, how much when what were independent of pricing and billing, build that artifact, get that instrumentation in place, deploy a cloud metering service. First, let that metering service surface, the data, the insights, and you’d be amazed what you may have other first thought of what your customer facing pricing plan would be what the data tells you will probably give you some far flung insights. And now on the backs of this data, you can now model what your pricing plan will be. So that is really the starting point. If you’re thinking about usage based pricing, deploy this technology stack of cloud metering, and you’re on the right
Swapnil Bhartiya 28:44 path with it. Thank you so much for joining me today. And of course not only talk about the company, but also the larger topic of cloud cost metering how to become more efficient. So thanks for sharing those insights. And I would love to have you back on the show because this is a topic which is kind of becoming a very big theme in the cloud at work. So I look forward to our next discussion. But I really appreciate your time today. Thank you.
Puneet Gupta 29:06 Likewise, I really enjoyed it. Thanks for having me. Yeah, I look forward to coming back.