We strongly believe that cloud computing has and will continue to have more impact than any other IT trend. Conversely, we think there’s a tremendous gap between the hype surrounding many new technologies/movements and their profit making potential. The hyped technologies include, but are not limited to “5G”, Internet of Things (IoT), Open Networking/SDN/SD-WAN/NFV, Virtual Reality, etc.
Cloud computing changes the networking industry in the following ways:
- The branch and spoke topology of private WANs and IP VPNs has been based on branch offices connecting to the centralized corporate computing data center (via IP VPNs, private lines, Ethernet virtual private lines, etc). That’s now giving way to branch offices accessing cloud computing resources via a variety of networking schemes that we’ve discussed in several previous articles (one we especially like is AT&Ts Netbond).
- The volume of networking equipment sold for enterprise/private networks is decreasing due to the flat to declining growth of premises data centers which are remotely accessed via switch/routers. Many of the mega cloud providers, like Google and Facebook, design their own switch/routers for use in their mega data centers. Other cloud players, like Microsoft, continue to buy switch/routers from legacy vendors (e.g. Cisco, Juniper and Arista Networks).
- Large backbone networks used to be designed and deployed ONLY by government agencies (like NSA, NSF, European Commission, etc) and global telcos with huge footprints. Those telcos include AT&T, NTT, BT, Verizon, Level 3, Deutsche Telekom, Telefonica, etc. Their backbone networks will not grow as fast in the future as none of those companies are large cloud computing or storage service providers. Even the US government is moving from premises based to cloud computing services, which decreases the need for government agency private networking.
- The big cloud players (Amazon, Google, Microsoft, Facebook, Alibaba, Tencent, etc) are constructing and adding on to their wide are backbone networks which inter-connects their mega data centers. For example, Amazon has recently completed deployment of a 8,700-mile undersea cable, which is part of its A.W.S. global network. Microsoft and Facebook are also building a large undersea cable network called MARIA.
- The control and analytics for the IoT are likely to be built on cloud platforms, such as AWS IoT. These are managed cloud platforms that lets connected devices easily and securely interact with cloud applications and other devices. AWS IoT can support billions of devices and trillions of messages, and can process and route those messages to AWS endpoints and to other devices reliably and securely.
- When standardized and certified “5G” service finally arrives (sometime in 2021 or later), we believe that remote user access to the cloud will be a very valid use case for IoT real time control/analytics and other applications that require low latency to/from wireless endpoint and the cloud.
NY Times article (Dec 26, 2016): Why the Computing Cloud Will Keep Growing and Growing, by QUENTIN HARDY
Jeff Bezos of Amazon, along with a couple of his rivals, may eventually control much of the $1 trillion global market for business computers and software.
That is because Amazon Web Services, his big-business computing division, is starting to affect more than just the world of computer servers, data storage and networking at the core of computing. Increasingly, it is also entangled with mobile phones, sensors and all sorts of other devices in the so-called Internet of Things.
It’s the same story at Microsoft Azure and Google Cloud Platform, the other two big cloud companies. Start-ups and giant corporations rent the core resources, along with related software, instead of owning and running their own machines.
What’s next? As innovations like artificial intelligence and connected devices become popular, customers are putting cloud components in mobile computing, home games and email marketing campaigns. In other words, the big clouds aim to be everywhere.
“When has Amazon ever thought about anything other than world domination?” said Lydia Leong, who follows cloud computing at Gartner. Not content to be in big centralized data centers, she said, “they want to be at the edges, whether that is a customer’s own computers or the Internet of Things.”
This aim for domination was clear at Amazon’s big customer conference, called Re:Invent, which was held in Las Vegas this month. About 32,000 people went to the fourth annual event.
In one talk at the conference, an Amazon Web Services executive showed off the company’s 8,700-mile undersea cable, part of an A.W.S. global network that each day adds computing power equal to that inside a Fortune 500 corporation, and spoke about this expansion. He talked about crushing the costs of servers and networking, most likely sad news for old tech giants that make those things, like Dell and Cisco.
In a nice bit of showmanship during the main keynote, Andy Jassy, the head of A.W.S., appeared onstage with an 18-wheel truck carrying a device that could suck 100 petabytes of data out of a customer’s computers and put it in the Amazon cloud. That is equal to two billion filing cabinets of paper, which a surprising number of companies now possess in digital form, thanks to things like video and sensors.
Put that together with some software Mr. Jassy talked about that would be on chips made by Intel but capable of gaining access to the A.W.S. cloud, and you get the picture: There isn’t a part of computing Amazon doesn’t want to touch.
It is easy to see why this matters to Amazon. In the third quarter, A.W.S. had revenue of $13 billion a year, growing at 55 percent annually. A.W.S. was 10 percent of Amazon’s revenue, but more than 100 percent of the company’s operating income. Amazon’s international retail business lost money, and United States retail sales are nowhere near as profitable.
Amazon says it is hardly moving away from a core business of providing large-scale computing, but rather finding more ways to sell stuff related to it by moving to edge devices.
“We see it less as a move from one to the other, and more of an extension,” an Amazon spokeswoman, Mary Camarata, wrote in an email Saturday. “We have an enormous number of customers excited about leveraging the capabilities.”
Amazon is not alone in this business, and the competition is getting more intense. A.W.S. now has 81 services, including ways to work on home video games. Microsoft’s 67 services include Internet of Things “hubs” and email marketing campaigns. Google has 53, including ways to deploy mobile software globally and steer performance with data analysis. Comparisons of services are difficult, as one company’s service may encompass two or three offered by another.
Machine learning — a method for computers to gain knowledge without being programmed with that information — is front and center for Alphabet’s Google, said Urs Hölzle, the head of technical infrastructure at Google Compute. Google has recently shown off its own global network of submarine cables, along with local devices like cloud-connected office whiteboards. Over the next year, Mr. Hölzle said, Google will open about one new Compute facility a month.
Building out across the globe, with sometimes $1 billion or more in a facility, is critical in some cases to meet local data regulations. Equally, the big cloud companies all want to be as close to customers and their devices as possible.
“Global proximity is a huge advantage,” said Corey Sanders, the director of program management at Azure. “This is a way to transform your business, including the way devices on the edge act.”
There are profound consequences from the scale and ambition of this trend. Given their size, wealth and technical expertise, the big cloud companies are likely to build cheaper designs and demand lower prices for everything in computing. Who is to say they don’t affect the devices themselves?
That is starting to dawn on the rest of the industry. On the first day of Re:Invent, Mr. Jassy had a private lunch with about 10 venture capitalists. It is an annual affair, where he indicates where A.W.S. is going, and they figure out how to make money from it.
“He wasn’t explicit, but if you were hoping to invest in storage, computing — anything below applications — you are hosed,” said Dharmesh Thakker, a partner at Battery Ventures, who attended the lunch. “Andy is smart and approachable, but reading between the lines, I’m not sure this is good for the V.C. ecosystem.”