Continued – Here are parts 3, 4 & 5 of Sramana Mitra’s, Silicon Valley tech industry Blog, Codero CEO and President, Emil Sayegh talks trends in cloud hosting.
Sramana Mitra: Give me the number again of where you would say the flip happens from a public cloud to a hybrid cloud.
Emil Sayegh: The flip is starting to happen as companies that started in the cloud are looking at their bills now and they’re seeing the exorbitant prices. I think there’re a couple examples. One of them is Target.com. They were with AWS. There’re a lot of public stories as to why they moved. There’s Moz. You may have read some of the publicity here in the last two months where they left AWS for pricing reasons.
This is the analogy I always give and it works. Essentially, if you have a constant workload for your application or even a workload that varies but in a rational or measured way between day to day and month to month, putting all your applications in the cloud is like living 365 days a year, 24 hours a day in a hotel. Sure, living in a hotel is convenient and you can scale up very quickly by adding more guest rooms if your family and friends started to show up. That’s not a way to live your life. It’s not that private. It’s expensive.
Sramana Mitra: So the main reason the flip is happening is that the price of hybrid cloud for really high workload, as companies scale, is much higher than what it would be in a hybrid environment?
Emil Sayegh: Exactly. With pure cloud, the price is much higher than it would be with a hybrid environment because hybrid environment allows you to offload your constant workloads on to dedicated servers. If you get a similar instance between dedicated servers and an Amazon Cloud Service, which have exactly the same specs, and you run them both for one month continuously, you’re going to pay three times more on AWS than on a dedicated server. For those workloads that are constant, on a hybrid cloud you can easily put it on a dedicated server while you leave your variable workloads on cloud.
Sramana Mitra: Can we talk about a different use case? Let’s talk about publishing. Do you have customers that are in publishing?
Emil Sayegh: We have customers throughout all the spectrum.
Sramana Mitra: With publishing, the customer behavior is a little bit different. Online video is a big trend. With a publisher that has enough video content with millions of people accessing that video content, what does the infrastructure requirement look like there?
Emil Sayegh: One of our big customers is one of the big networks. What they have on our array of servers is a whole bunch of videos for their affiliates. All the videos that their affiliates show on their affiliate websites are all served up on servers that are at Codero. That’s one of the examples. Those guys are using dedicated servers and not cloud because of performance consideration. You have to render that video. That takes a lot of processing power. They are using an array of dedicated servers and then the front-end interface of each one of the customized web interfaces to their affiliates is put on cloud. But the hardcore data storage and data rendering is all done on an array of dedicated servers on a hybrid type of structure.
Sramana Mitra: Is there anything substantially different in any other segment of customers that you cater to besides the two use cases we discussed?
Emil Sayegh: You have SaaS customers that are similar to the e-commerce value proposition.
Sramana Mitra: These are SaaS customers who are themselves public cloud vendors?
Emil Sayegh: They are public SaaS vendors. They’re leveraging our cloud infrastructure to offer SaaS to consumers.
Sramana Mitra: What are we seeing in that segment? That’s an interesting segment. I’m sure you have a customer base in that segment.
Emil Sayegh: Yes, we have a great customer base in that segment. They’re very similar to an e-commerce customer from the fact that if their application is down, their whole business is down. We cater to companies that make their living off their web presence. SaaS companies’ value proposition is very similar to the e-commerce customers. From an architecture standpoint, they are also similar because a SaaS company needs to scale up very quickly at the front-end with as many users that are logged in at that point. They need to create as many of those sections as they can.
I’m going to give you this odd example. We have one company that provides SaaS to agencies that are involved in rescue efforts and act in times of disaster. When there are no disasters in the United States or in the world, they’re probably not very busy. But as soon as there is a disaster, then you have a bunch of agencies that are descending on one region and they need to be able to process the cases. It’s kind of like Salesforce on steroids but focused on disaster recovery. That’s their specialty. In terms of need, they scale up extremely quickly. They spin up a bunch of cloud instances in the front-end. They may add a couple of database servers in the back-end so that they can accommodate that load.
This is an unusual example that I’d like to give for SaaS companies that benefit from the hybrid infrastructure because the corollary of this is if you were to go with a monolithic, either cloud or dedicated, they would have to essentially commit and get a bunch of dedicated servers – if they want the performance – or they had to just get stuck with just cloud and suffer from a performance standpoint. When a disaster happens, they may clobber their database servers and their database servers may stop working in the most immediate time of need. If they get a database server that’s dedicated and they get a bunch of cloud servers that are just invoked at time of need, that is the optimal, most economical, and highest performance solution that they can have.
Sramana Mitra: I’ll switch gears a bit. Given all these things that you’re seeing in the market, can you look ahead to the 5 to 10 year horizon. Tell me your thoughts about what’s going to happen. What do you anticipate as new that is going to happen?
Emil Sayegh: We are all heading toward a world where data is being collected every micro second, where wearable devices are going to be common and where data is collected even in our house. Vital data collected from us can be predictive. We’re heading toward a world where massive amount of data is going to be collected. We think we’re collecting a lot of data now. With avant-garde customers that I’m talking to, that data is going to take shape in the form of numbers, videos, and pictures. To accommodate this huge flood of exponential growth in data, we’re going to have to find ways to store and retrieve data very quickly and very efficiently. That’s going to be the next frontier for a lot of us. That’s going to impact storage – the cost and performance of storage.
Also, a lot of what I’m talking about here is going to be bottlenecked by how big and how fat their network pipes are. You’re going to start to see winners and losers in this space of accommodating all these data based on how big the pipes are and how inexpensive their storage is. Kind of like a price and performance ratio on storage more so than it is about pure computing performance like what we’ve seen to this point. I would say over the next ten years, storage is going to be important. High performance networks are going to be much more important. Of course, you said that hybrid is going to be much more prevalent and important. What we’re seeing is an exponential growth in the amount of data storage and retrieval and how efficiently you can do these things. Those two things really hinge on storage and network performance.
Sramana Mitra: Given that discussion as a backdrop, where would you advise young entrepreneurs to look for opportunities to build new companies?
Emil Sayegh: Infrastructure as a service is yesterday’s battle. You have companies that are well established. What you do with that infrastructure is the key. I see young entrepreneurs focusing on how to use that amazing array of computer resources that’s available to them and providing services that we haven’t thought about yet. Let’s take the example of medical service. If we’re sick, we’re in the hospital. They monitor our vital signs. What if young entrepreneurs really focus on the predictive nature and utilize the vast amount of infrastructure, computing, and storage resources that are available out there. I would advise them to go one layer deeper into yesterday’s opportunities and look at how how efficiently you can do these things to build new companies.
With all these vast amount of data available, how can we come up with algorithms that will give us the ability to extract quickly or organize data quickly so that we’re utilizing this massive amount of data and organizing it in a logical way. I see opportunities at the usage and at the utilization of data more efficiently.
Sramana Mitra: Thank you for your time.