You are here

Cloud Computing & Machine Learning: Past, Present, and Future

Presented by Rich Rao, Global Head of Devices and Education, Google

Willie Sutton was an American bank robber who stole an estimated $2 million during a forty-year criminal career. When a reporter asked why he robbed banks, he answered: “Because that’s where the money is.” ‘Sutton’s law,’ as it is now called, can be applied to your business too. When contemplating investments, you should go where the money is. And in today’s environment, that means investing in technology, and more specifically, the cloud.

In my ten year career at Google, I have helped large companies, small businesses, and educational entities implement technology. The experiences have given me a glimpse of the past, present, and future of cloud computing and machine learning.

The Past

They say that necessity is the mother of invention and that was indeed the case for Google. It was 2007-2008, during the global financial recession, when Google announced that they would be launching their own browser. People wondered why, saying that there were already enough browsers on the market. However, people at Google were noticing several problems that would be solved if they created their own browser. For one, the internet was not as fast and instantaneous as it could be. It also wasn’t very secure, with more and more security breaches happening through browsers. Lastly, Google was stalled with the current browser options. They couldn’t build the apps they wanted or do rich things with client-side technology.

Flash forward to today and Chrome is the leading platform on the web, but more importantly, it pushed the paradigm, transitioning browsers to cloud computing. Several cloud advantages became clear, including: 

●        Multitenancy: With the cloud, multiple people are housed under one roof, so to speak, which provides better resource utilization. Even better, it means that software is always up-to-date, with changes happening behind the scenes in real time and no upgrades or new versions of software necessary.

●        Near Infinite Scale: With a near infinite scale, you can get elastic computing. However, even better is that costs decrease over time or you get more things included for the same price point

●        High Reliability: When you move toward professional management of software and hardware, you get better reliability which means high uptime. Even better, there is no scheduled downtime, which is beneficial because downtime can cause enormous losses for users.

●        Automatic Redundancy: With the cloud, you get backup, which can help your business recover from local disasters (like hurricanes or power outages). Even better, by placing data across multiple data centers, you get advanced security through encryption.

The Present

Google has expanded its cloud platforms by creating technology that has solved problems for the company. These technological breakthroughs have allowed Google to scale, and then the cloud platforms are offered as open source technology that can help other businesses grow and solve their own unique problems. This includes:

●        Management: Google Stackdriver, Identity and Access Management (Cloud IAM), Support

●        Compute: Compute Engine, Preemptible VMs, Custom Machine Types, App Engine, Container Engine

●        Storage: Cloud Storage, Nearline, Cloud SQL, Datastore, Bigtable

●        Networking: Virtual Network, Load Balancing, CDN, DNS, Interconnect

●        Data: BigQuery, Dataflow, Dataproc, Pub/Sub

●        Machine Learning: Cloud Machine Learning, Speech API, Vision API, Translate API

Some examples of the ways companies have used these Google cloud platforms include:

●        Snapchat: Snapchat is a photo sharing and messaging app. Although it only started in 2011, it is now one of the top ten most downloaded apps and it consumes 1 percent of total internet traffic. The founders used App Engine to build the app. Their team has no infrastructure engineers, they simply conceived the app and figured out how to monetize it – a business model for the new era. 

●        Zulily: Zulily is a company that sells clothing for the family. They started just a few years ago and are now at a billion dollars in revenue. One way they have gotten ahead is by providing specific promotions and offers based on customers’ preferences and past purchases. They were challenged because they had both structured and unstructured data that didn’t work well together, but then they began using Hadoop and BigQuery, which allowed them to analyze all the customer data to provide personalized and targeted offers.

●        SunGard: SunGard is a financial service audit company that records all the transactions that happen in the market. With six billion market events every hour, there is a lot of data to sift through. However, using Cloud Bigtable they can process and run queries on the data, with the ability to identify any transaction in the last six years in just seconds.

The Future

The technology research and advisory company, Gartner, predicts that by 2018, 20 percent of business content will be authored by machines and by 2020, 40 percent of mobile interactions will be facilitated by smart agents. Google will continue to explore cloud technology and machine learning, including making Google Apps like Gmail, Docs, and Hangouts more intelligent. Google Search will include speech recognition, Gmail will get a smart reply feature, Google Photos will get a more sophisticated search function, and so much more.

As we move forward with cloud and machine learning technology, the goal will be to move beyond productivity to make applications and processes more natural, more automatic, and more predictive.

Today you can build an app for just about anything, and you can use the cloud and machine learning to build a stronger business. I suggest:

1. Do the easy stuff first. Are there some easy wins in terms of SaaS for your end users and IT?

2. Join the culture of innovation. Now is the opportunity to build exactly what you want to build. What problems could you solve if you had no technical bounds?

3. Pay attention when recruiting new team members. All major universities are now teaching about topics like cloud platforms and machine learning API. Think about that when you recruit new members, and also how your team can grow skills internally.

In my career, I have worked with large companies, small businesses, and educational entities. In the process, I have noticed that while enterprises large and small are trying to figure out how to move to the cloud, many schools have already done it. This is ironic because they typically don’t have the largest budgets or a large tech staff, but they are finding a way to get it done. This demonstrates that anything is possible for your business too, despite any limitations that you think you have.

 “We launched a browser, and it turned out to be a pretty good one, and now we have people using it. It’s the leading platform on the web, but more importantly, it’s pushed the paradigm. This is an example of the transition towards cloud computing when it comes to browsers.”

“Instead of us working for the machines, the machines can actually work for us.”

“All you have to worry about now is the language and can I build an app, then everything else can be managed for you. What this means and the reason why this is so impactful to businesses is you can just build an app for almost everything. And it’s happening today.”