Google's thoughts on the direction the cloud is drifting towards
Rick Harshman is the Managing Director of Google's Cloud business across Asia Pacific and Japan. He is responsible for all aspects of Google Cloud Platform's revenue attainment, GTM, and strategy across the APJ region. In this interview with Enterprise Innovation, Harshman speaks of the direction the cloud is heading and the role Google is looking to play in this space.
Could you take us through some of Google Cloud’s latest developments in Asia?
This isn’t just for Google Cloud, but Google in general—there’s a big level of focus on Asia from the infrastructure perspective. We’re building out cloud platform regions in various countries in Asia Pacific—we just launched Japan, the second behind Taiwan, and we’ve pre-announced three additional regions for Asia Pacific that will come online in the next 12 months—Singapore, Sydney, and Mumbai. So far that’s what we’ve announced for 2017, but the nice thing is we’re not done. There’s a number of conversations going on about additional locations as well.
This is driven by demand from our customers, who have been very excited about leveraging the power of Google’s cloud services. We’re layering in those infrastructure investments on top of our people investments because our customers need low latency and very strong performance from a location perspective. We’re making investments in some specific areas. The first thing is called the Advanced Solutions Lab, which we started first in the Valley, but we’re seeing tremendous interest from Asian customers in it as well. What’s really unique and interesting about this is that companies bring themselves and their ideas to Mountain View to work with our machine learning experts. We’ve got a backlog of companies that are interested in engaging in this because machine learning is an area where Google has extremely strong depth in. We saw an opportunity to externalize the machine learning expertise we use for our own services, and the Advanced Solution Lab is a step further to really engage companies on their ongoing projects.
Is there any concerted effort in Google’s part to develop in the future an R&D base in this region?
Typically, multinationals—specifically from the West—would have a sales and marketing base in Asia and that’s about it. You’re right about the R&D—that’s where Google is seeking to make a big difference with its Next Billion Users initiative. It’s not specific to cloud, but what I’m really happy to see—and this is something I haven’t seen in my career—is that we’re having a number of our engineers continually coming out here in the region to meet and get feedback from our customers. These are not pre-sales engineers, but the actual folks who build our products and services. It’s the third time they’ve come out here, and the level of conversation and the depth they go into has made all the difference. We’re seeing not only light bulbs going off on both sides, but also the level of relationship building that comes out.
In the age of infinite customization and granularity, that engineering to end-user link is quite critical.
Most folks believe Asia is just cost-sensitive but it’s not. They want to be able to see the value, the technical benefits they’re receiving as they meet their financial requirements.
One thing I want to mention is that we’re really focused on being seen as the open cloud option from a public cloud perspective. That’s important because if you take TensorFlow for example, we created it a little more than a year ago and then we open sourced it and gave it to the community. What’s interesting about this is it provides companies the ability to take their data and create machine learning models on their data. Of course we think that you’re going to be able to run TensorFlow best on Google Cloud, but we’re not just keeping it for ourselves—we’ve opened it up, it can run on other public clouds, it can run on private installations and hybrid environments, and that’s something we’ve seen really pick up.
Do you think that open source model is the future and that the closed-off systems are going to be the dinosaurs?
I think what you’ll see is that the companies that listen to their customers the most and continue to be open and flexible, will be the companies that continue to be successful. Take for example Hadoop. That overall ecosystem came out of Google, and it’s now a multi-billion dollar ecosystem across the board. I think you’re going to see that type continue to bloom with things like Kubernetes, which was just open sourced a little more than a year ago. It’s crazy, the pace of innovation today. That’s what enterprises are really looking for. If you’re an organization that can help a business get faster time to market, get greater agility and flexibility at a lower cost with no lock-in from a contractual perspective, you’re going to have a pretty compelling offering—especially to an enterprise.
For Google’s part, we’ve been running our own cloud service internally for 15 years. The difference now is we’ve decided to externalize those services and make them readily available for any customer via APIs and have them leverage it in their business. We’re seeing a lot of interest from industries and a wide variety of use cases—from basic things like dev test on websites and mobile apps, to more advanced solutions around managed data analytics and machine learning. What’s interesting is where companies begin to leverage cloud to store their data and then get greater insights to that data by leveraging these analytic tools and services and get more actionable near-realtime insights. Layering that on top of our machine learning attributes not only answers questions to problems they’ve had, but gets them looking at what new models they can create that could help them gain greater efficiencies, higher revenue, and so on. That’s the area we’re most excited about—helping potentially reshape the way organizations think.
How do you see Google Cloud evolving?
If you want to be considered a player in this market, you have to have the underlying base services that you mentioned—compute, storage, database, networking. I think the differentiators across the board come in the following areas:
Cutting edge technology. When I talk about cutting edge technology, we’re talking about focusing on machine learning and the data analytics pieces. We feel we have very strong offerings in both areas and will continue to see an increase pace in innovation around those areas.
An unparalleled high performance network. We're making a lot of investment on the networking side specifically this year, one we already launched, which is the faster network that connects the U.S. West Coast to Tokyo, and that’s at 60 terabytes per second. That connects onwards to Taipei, which is an additional 20 terabytes per second. We’ve also pre-announced that we’re building the Pacific Light Cable Network (PLCN) with Facebook—what’s amazing about that is that it’s an additional 120 terabytes per second. These are undersea cables we’re running from the U.S. West Coast to Hong Kong, and that is to provide better performance and connectivity for Google services. That is a huge differentiator because when you are using Google Cloud and our services, you do not go on the public internet. You remain on the Google backbone, and that is a very compelling offering for companies from a performance perspective.
Security and reliability. These are our top two priorities if you think about how we operate. If we don’t have extremely strong security and highest levels of reliability, we’re not going to have a strong business because people just won’t trust us. From a security perspective, just for Google Cloud we have over 600 security experts working on ensuring strong levels of security throughout all layers of the stack. There are also folks working on certifications to be able to meet regulatory requirements that end customers need to run on the platform, so we have those teams as well.
Openness. We’re going to continue to develop products and services that we’re not just going to keep for ourselves but we’re going to open source. The two most popular GitHub repositories right now are Kubernetes and TensorFlow, and they both came from Google. That’s pretty cool to see especially since they’re both only a little more than a year old.
Finally, our customer partnership model. It’s not going to be something where it’s customer-vendor and you’re never going to hear from us again. Our overall commercial philosophy is ‘no contracts, no commitments, and per-minute billing’. That is such a huge differentiator. An average enterprise has around 500 to a thousand workloads in their environment. If you can look at each one and you can determine via optimization automation what you need to run for 24 minutes on an hourly basis as opposed to a full hour, that’s a substantial amount of savings when you multiply that across your entire fleet.
There’s this entire spectrum of sentiments from cloud is good to cloud is evil—where do you see the sentiments of some of your customers and potential customers here in Asia falling under?
Being on the front lines over the last seven years, what I’m seeing is the level of skepticism in perception that the cloud is not going to be as secure as what you can do on-premise yourself, has dramatically changed. When I was at a former company back in 2009-2010, it was all “Cloud? What’s that? I’m not going to do that.” Security is still an area that companies still have the responsibility to ask about and do due diligence on, but is it something I hear as the first concern today? No. Now there’s been an overwhelming amount of education in the market.
When it comes to security, ongoing education still needs to be done because if any cloud provider has an issue with security, everyone gets painted with that brush you’ve mentioned. With that being said, specific to Google, we invest a lot in building the most scalable and secure services that we can provide. But security is still a shared responsibility with the customer. We can give them recommendations, can encrypt their data in transit and at rest, but if they don’t take recommendations on how to best architect with security in mind, and if they accidentally leave a port open as an example, well there’s things you can absolutely lockdown and that’s where the partnerships with our engineers come into play.
What excites you about where the cloud is going to move towards the future and what the cloud is going to enable in the future?
I think what’s really exciting is that the cloud is rapidly becoming the new normal. When companies are coming out with new products and services, they’re looking at cloud-native applications. I think what you’re going to see is Asia moving from a mobile-first environment to an A.I.-first environment or machine learning-first. That is where you’re going to see a tremendous amount of innovation for individual companies and verticals. You’re going to see use cases that people haven’t even thought about yet. Today there’s dynamic pricing, recommendation engines, credit risk analysis—things that when you can apply machine learning models can create greater levels of efficiencies for you and lower your costs—but these are all still fairly basic. What’s going to be really exciting are what you’ll see in the next 24 months. Technologies are accelerating quickly and I think a lot of that innovation’s going to come from Asia. It’s a very youthful demographic and it's already a mobile-first environment - how does that evolve and accelerate?