During a presentation at the Cloud Foundry Summit, Siki Giunta (SVP, Verizon Cloud) outlined how cloud technologies can drive enterprise transformation. She discussed the application portfolio framework routinely used by enterprises to assess workloads that need to be “killed”, “maintained” or “transformed”. Enterprises currently allocate the bulk of resources—budget and personnel—to maintenance projects, and only a small percentage of resources to transformational projects. Consequently, they are beleaguered by sluggish innovation and missed market opportunities.
According to Giunta, cloud technologies, and platform-as-a-service (PaaS) in particular, can significantly address this state of affairs. Whereas infrastructure-as-a-service (IaaS) can incrementally lower costs and enhance margins, PaaS enables disruptive innovation that leads to enterprise transformation and topline revenue.
Giunta dispelled the notion that individual technologies such as containers or Puppet/Chef can provide a meaningful application-level orchestration platform. She argued that a full-featured PaaS is preferable, because it weaves various technologies into a cohesive and automated platform that lowers the burden of application maintenance. As a result, a PaaS can free up both budget and personnel for more transformative projects which target new revenue opportunities.
Further, a PaaS can also power products and services aimed at these opportunities. Capitalizing on new revenue opportunities, which may be fleeting, requires an enterprise to deliver products and services with agility and scale. A PaaS enables such capabilities. In fact, Giunta predicts service providers will find it difficult to attract customers without a compelling PaaS offering.
The talk concluded with a brief review of the success Verizon achieved with Pivotal Cloud Foundry. The Verizon Cloud Team successfully employed Pivotal Cloud Foundry to rapidly convert a mobile application for Verizon technicians. The previous version of the application was difficult to scale and based on an expensive, traditional software stack. Pivotal Cloud Foundry enabled an application that was highly scalable and used cost-effective OSS technologies. Building on this momentum, Verizon hopes to pursue more complex Internet of Things (IoT) projects with Pivotal. “The next big thing that we’re going to do is obviously see if Pivotal can help us with a PaaS for IoT,” Giunta said.
Watch the full talk from the Cloud Foundry Summit:
- Missed Cloud Foundry Summit? Read the recap here.
- Watch more Pivotal sessions.
- See all sessions from the Cloud Foundry Summit 2015.
- Learn more about Pivotal Cloud Foundry.
What I’m going to do today is give you an idea of why large enterprises, medium enterprises decide to transform and the path to innovation and the cost of innovations and our experience at Verizon.
We’ll talk about the case. We’ll about a little bit of one concept that I’ve developed for the longest time that’s called the Application Portfolio Value so that you understand how enterprise IT thinks about it. We’ll talk about IaaS and PaaS and some examples.
Enterprises, this is the way that they normally have their resources allocated. Resources could be the magnitude of dollars, the people, the very cool geeky people, the not very cool geeky accountant, but that’s how they do.
Seventy percent, they keep the light out. If they’re very innovative, maybe 20% they do it and 10% compliance. Compliance now is growing a little bit more. That’s why a lot of enterprises, they are slow to market. Really PaaS what would allow them is to reduce the cost of maintenance of the applications and to keep the lights on and to move all the dollar, beautiful people, amazing brains to innovations so that they could provide better services to their business. That’s really all about how can I do it better, faster sometimes with less people but maybe different types of people.
Let’s do a little bit of math. When you do cloud or actually every enterprise, every two or three years to do a portfolio analysis, meaning I look at all my applications and I look at the one that I want to take forward, the one that I need to build and the one that I have to retire. There is this application portfolio evaluations that is the functions of as a service strategy and you divide your portfolio in a Kill, and by way it’s called “Kill”—meaning I want to decommission, because you got to kill something to move the resources somewhere else. Things that I need to maintain and obviously how much more I can transform.
If you do just IaaS, normally you will be able to kill 25% of your capacity, of workload capacity. Some of this very old mainframe things, otherwise, you have to find a way to move them. I’ve seen a couple go to Java very nicely, but it’s a bit of a hard game.
Isn’t it easier to write a brand new app? Why don’t you kill some of these applications? The biggest problem is maintenance. Maintenance is still 60%. You’ll find the guy in there saying “I want my report. I want to keep it there” and you keep saying I know, I have to kill this report, I have to kill this application so I can build a more transformation. Transformations become and produce only 15% of it.
If you put PaaS in there, you see killing is still the same. More or less, it’s an evolutional service for a company but the maintenance gets reduced and that dollar goes straight to transformation and innovations. If we combine and I think that’s really my case. You have to combine infrastructure as a service and PaaS service to get more ability to kill and decommission services. To get to really ideally only 25% all the time to maintain and to move a bunch of dollars and people to your transformations.
Only using tools as a service, moving opex and capex, you can actually move the dynamics of this transformation. By the way, if you go to any enterprises and you ask them about their KMT most of them they do have it. They keep it secret to you because they don’t want know how much you could make the life difference but really it has to be there. Otherwise, corporations become obsolete.
Let’s talk about IaaS and PaaS. By the way, I run a great development team for cloud. Our team is the commercial, meaning selling cloud services to end user and Verizon IT being customer one Verizon IT. I have a great amount of developers that really like building things from scratch. They really bring all their skills of “Hi, I am a network engineer”, “I’m a storage engineer”, “I am a computer engineer” and they just love doing that, you know, building it.
It has a certain fascination about when you start getting to scaling and global deployment then our job has been “Why don’t you use a PaaS service to build a service catalog of some services that go faster.” Instead of you having to rebuild again API, services and everything, why don’t you just containerize, you make it as available distributed services so we can go much faster. Because in cloud is all about faster, continuous development, continuous innovations.
Let me give you another idea. Why infrastructure as a service is not enough and why at the same if you talk to an enterprise, everybody thinks that virtualization is infrastructure as a service. I thought that in 2010 we demystified that. No, I still go there and they think that they’re doing infrastructure as a service. Infrastructure is not enough. You get a better opportunity cost, you move capex, opex, you have better utilization of your resources but your revenue component doesn’t accelerate that much.
With the PaaS you have the ability to work on the two dynamics. You can reduce your cost, opportunity cost on maintenance and use IaaS and the opex, capex situation. But the beauty of it—you affect the revenue. I’ll give you an example. When you can go to a large CIO and tell him that in less than four months he can have a brand new airline kiosk activations in a major airline he will listen. He will listen straight away. The only way that you can do that is using PaaS and PaaS will activate the workload, because the most important thing is IaaS is infrastructural, PaaS deals with workloads, the creations of workloads that generate the business and they produce the revenue.
The other big, I think, misconception in my point of view is that a lot of people think that Puppet and Chef is PaaS. Sorry guys, it’s not. It is PaaS for infrastructure but still deals with infrastructure component. You can use it, you can make your life easier, you can scale better. As a service provider, I adore it.
I make money if I can maximize the cluster usage because I only make money if my clusters are used more than 30%. Below 30%, I lose money every single time. When I get to 70, 75 and I can move fast, mucho money. Meaning, dollars and margins are 30, 40%. I adore these tools. I’ll tell you about container in a minute. But, that will not allow more workload to jump on my clouds. It will make my clouds more efficient. Great, I can make more money but I will not have customers to very readily fast deploy application into my cloud. That’s why you have to have PaaS services.
Our experience with Pivotal has been very crucial. Because we were able to show to take in a traditional, standard stack applications even cool Java application and decided to take it away from a traditional stack to an open source stack and move very fast to go to the next level that I think is the mother load of PaaS and that’s mobility.
I talk to a lot of customers that love Kony for mobility. I say to them all the time, your path to fast mobility has a stop into the PaaS. It’s like you can’t really build mobility applications that don’t run in a cloud because it doesn’t make sense to accelerate that. You could accelerate it faster if you actually create your application logic in a PaaS environment.
We all want to go to mobility because for employee optimization, mobility is the only platform. Today, an employee IT per year cost four and a half thousand dollars. You got to get them their PC, got to be somebody there. What about if I can take 35% of my people and just give them a tablet and they can do everything in there, the cost drops below $800. That’s why corporation wants to move fast optimization of mobility but if you don’t have a dynamic PaaS environment on the back the journey is much longer.
I think a lot of people try to go from IaaS to PaaS to a mobility platform and it really doesn’t work for scale. It doesn’t work for hyper scale and reusability and your ability to learn while you build your applications. Let’s talk about developing in a corporation. A corporation, they’re still doing a lot of waterfall.
You remember this maintain piece. That maintain piece is 100% waterfall. Think about how slow and painful it is and the move to agility or continuous deployment is very, very hard. We launched a brand new cloud services in October and we were in this nightmare of waterscrumfall and it was just slow. Even if we were doing it on a monthly basis, so we pushed it to a continuous development on a weekly basis.
That was only using PaaS and having the ability to accelerate a continuous innovation and beat the curve. I’ll talk about containers. I love containers. It’s painting by numbers sometimes. It’s really incredible. I use a little bit of this, a little bit of that. I put it together and things move around. I put in some nice property and I tell him go faster, faster, faster. Go over there. Don’t go over there.
The security is not there. Large enterprises will hate it. We only today would use containers on dedicated clusters, because in a multi-tenant environment, when you specially use micro services where you have loosely coupled objects, these objects get themselves in trouble spots. We cannot control it.
If you put monitoring or dependency mapping in a container you lose the agility of what they have. I love them. I think PaaS and container are the best way to write modern applications but it’s going to be a little bit of time for us to find how to harvest the security in a different way. There’s a couple of companies that they’re trying to really not for us. Enterprises adoption of cloud now is totally connected and linked to security and how we can implement it.
Continuous innovation, freedom, ability to build the most beautiful fast running applications but we have to do it in a secure way so that enterprises will continuously adapt PaaS and cloud technology. Why do you need both? As I said, IaaS affect your margin. The more I can condensate, the more I can do cluster management, the more I can cluster together the better margin but PaaS, it gives you deep impact to revenue.
I tell you, that’s a very important concept. A lot of time it’s lost. A lot of people think about IaaS as hyper scale, autoscale and everything and that’s cloud. I think cloud is big “S” and small “I” and you can only affect the big S, meaning service with workloads and PaaS. We have to use both as cloud providers, because our user will demand both.
The other things is we need automations. Clouds will only grow making money if you’re automated. The less fingers on a cloud the better it is. You’ll be amazed how much still finger are in the cloud and in operational cloud. Elasticity, especially if you want to go across zones, meaning your data center, not a data center, not a region or another data center.
If you have the more compact the workload is the easier it is to be able to move it for margin and that’s what it’s all about. Revenue, PaaS brings me revenue, automation of PaaS. IaaS brings in margin. The other things is about scaling. I learned about scale at Verizon. Verizon, it’s $140 billion, 150,000 employee, many countries, so when you do one thing they cannot do it just for a group of family of people. They have to do it at scale. You’ll be surprised how much applications that everybody thinks are going to scale in the real world of cloud, they actually do fail.
They fail for three reasons. First of all, they become impossible to operate. You go back to the fingering, you go back to I’m going to optimize this, a little bit. I’m going to put all this nice scripts in there and that’s impossible operation. Second, they really don’t run because you’ll have to say what about China, what about this country, what about that country. You have to embed and build services that are available.
Third, they really don’t make money if you don’t scale. Washmycar.com in Cincinnati is just one Washmycar.com in Cincinnati. What about if you do washmycar.com around the world that you can very fastly find the car wash that you like and it’s free now and you can wash your car and you actually get a coupon for 10%? These kind of applications needs scale and you can only build that with a PaaS.
What did we do at Verizon? We did a pilot and it was a combination of Verizon IT, to me, customer #1, and our cloud team. What we did, we took an application called Tech Tablet. This is the applications that the fires engineer use to schedule workload, to optimize the workload, to get data, to understand “Do I cut this wire or that wire?” It’s really an important optimization application.
It was kind of born for the cloud but “ish” as I say. I think it was thought that it would run—it runs on a cloud, it’s a virtualized application. But, can they push it out to everybody around the United States? What we decided to do and it was very expensive. The components of the applications were traditional software component.
Goal number one, can we move more applications to an open source environment? Goal number two, can we do it very fast? Goal number three, can we scale to one to three instances at speed? It was very successful. We were able to prove all three use cases and we gained both. IT gained the ability to move more mobility applications and us in the cloud environment, we gained a very strong use case scenarios, a very strong platform where we can actually take out—we are planning to take it out—in the next two months as platform as a service for our users. They can buy directly from our council and everything.
The next big thing that we’re going to do is obviously see if Pivotal can help us with a PaaS for IoT. As you know, IoT is another one of these incredible application development challenges where the stack gets even bigger and the dimensions get big. In IoT you have 5 dimensions. You have connectivity. You have machines. You have big data. You have big power computing data. You have human interactions.
These five elements of IoT—without a strong PaaS writing application for IoT and integrating it with our writing services that you can reuse—is a challenge that we cannot really take on. At Verizon, we are all about IoT and IoT is all about what’s going to happen in the future.
Thank you. That’s my story to you today. Have a great conference.
About the Author