ARCHITECHT Daily: Cloudera's IPO will be an eye-opener on open source

Cloudera officially filed its S-1 report on Friday, in preparation for an IPO later this year. You ca
ARCHITECHT Daily: Cloudera's IPO will be an eye-opener on open source
By ARCHITECHT • Issue #51
Cloudera officially filed its S-1 report on Friday, in preparation for an IPO later this year. You can get a high-level overview of its filing and the IPO market in various places. Here are the financial highlights, from a CNBC story
  • What they’re hoping to raise: up to $200 million
  • Revenue: $261.0 million in the year ending Jan. 31, up from $166.0 million a year ago
  • Net loss: $187.32 million, narrower than the $203.14 million from a year ago
Of course, if you’re really interested in the economics of big data, and Cloudera specifically, you should just read (or skim) the whole thing.
Cloudera’s S-1 lists a goldmine of open-source-related issues—some more worrisome than others—including the pace/continuation of open source development, licensing issues, and potential problems with the Apache Software Foundation. But I think the biggest thing to watch over the next couple of years will be broader than just Cloudera. Namely, it will be the effect of more publicly available financial data—especially from a high-profile company like Cloudera—on open source business models, whether or not companies are selling Hadoop. 
Red Hat aside, many open source companies still struggle with questions like finding the right balance between open and proprietary, balancing licensing revenue against support revenue, and balancing internal R&D investment with community investment. More evidence of what’s working and what’s not could be very useful for open source companies and investors alike.
And although this will ultimately matter across software sectors (and although Cloudera lately cites IBM as its primary competitor), the biggest immediate impact will likely be on Cloudera’s legacy Hadoop-based peers MapR and Hortonworks. I’ll ignore MapR for the time being because it’s still a private company with private financials, and I really don’t know what they look like. Hortonworks, however, has always taken a more traditional support-and-services approach to open source than either Cloudera or MapR. 
And it has experienced, shall we say, a roller-coaster ride since going public a couple years ago—in part because of that pure open-source product strategy. To its credit, Hortonworks isn’t oblivious, and has been doing some interesting things around IoT (with its DataFlow offering) and cloud computing (with its Data Cloud for AWS offering) that could help out. The company briefed me on them last week, and perhaps I’ll write up a little more on this shortly. 
I made a little chart (below) that’s simple in terms of metrics, but that highlights a very important distinction: While revenue is growing nicely at both companies, Cloudera’s net loss was smaller and shrunk during its last fiscal year. Hortonworks’ net loss was larger and grew during its last fiscal year. If I’m a Hortonworks executive or shareholder, I’m looking at that and thinking of ways to narrow the gap.
Basically, though, we’re in a period where open source is the norm for enterprise software (especially at the infrastructure level) and where SaaS and cloud services are catching on very quickly. Sunlight might be the best disinfectant, but it’s also good for illumination. As more open source and cloud-based companies go public (hello, Dropbox, in a bit!), we’re all going to learn a lot about the economics of modern enterprise software companies—the good, the bad and the ugly.

If you enjoy the newsletter, please help spread the word via Twitter, or however else you see fit.
If you’re interested in sponsoring, please drop me a line.
Use Feedly? Get ARCHITECHT via RSS here:
What's new on ARCHITECHT
Pivotal CEO Rob Mee on the success of Cloud Foundry and learning to sell enterprise software
Pivotal CEO Rob Mee discusses the success of Cloud Foundry in the face of lots of competition. He also discusses the shift from partnerning with startups to selling to enterprise software.
Artificial intelligence
This is a remarkably frank assessment from the founders of Skymind on how enterprises and startups should think about AI—how to build something, and whether they should even try.
Gary Marcus has been talking a bit lately, since leaving his post at head of Uber’s AI lab. This is a good Q&A about why Google, for example, has embraced deep learning, and where he would invest billions of dollars.
Maybe, maybe not. But here’s more than 130 companies arguably doing AI, grouped into the markets where they sell.
Good advice from Proctor & Gamble and American Express on how they are approaching AI. Some of this mirrors what you hear from other large enterprises, although I’m not sure how many others are willing to make such large internal investments.  •  Share
The Open Philanthropy Project is joining the entrepreneurs and organizations helping back OpenAI’s mission of safely developing AI technologies.
The author’s “change the locks” analogy is pretty nice. Essentially, it’s easier for criminals to break into your company if you’re using the same software as everyone else.
This is similar to my sentiment when bought Kaggle, which is that Google now has an inside track to influence AI developers and learn what they like.
Baidu announced DeepVoice a few weeks ago. Here is Google’s take on AI systems that can speak like real people.  •  Share
Cloud and infrastructure
Despite all the innovation in data centers over the past several years—from server designs to building designs to where they’re located—there’s still a lot of work to do. And while websites and cloud services have a global reach, the effects of the data centers that power them are often felt locally. Here are three recent stories that highlight why web/cloud infrastructure is still an unsolved problem:
This isn’t the biggest-deal hire you’re going to find, but it underscores Facebook’s continued ability to attract top talent, as well as its commitment to ramping up performance while minimizing resource usage.
It’s just a rumor, but the idea of Oracle buying Accenture would actually make a lot of sense, including as a means through which to push Oracle’s cloud services. But both companies deny any talks. And the price tag would be ENORMOUS.  •  Share
Says CEO Taylor Rhodes: “We thought the world wanted another alternative to public cloud. What we are learning is the world doesn’t need another public cloud, so OpenStack is shifting form and going private cloud.” I would focus on hybrid and not OpenStack.
Regardless whose cloud you’re running on, or what you’re selling, this is just solid advice. It’s also a lot easier to pull off now thanks to advances in software development, including container technologies.
So says a group of researchers from USC. Cloud brokers and other utility-type models have been suggested before, but perhaps now there’s financial incentive to actually do it.  •  Share
All things data
These are two good posts on the relative importance of data visualization, which is simultaneously useful, hard to do right, and of indeterminate importance to companies. I would suggest it’s most important for companies (e.g., news outlets) serving consumers with, essentially, one-shot charts. For many BI use cases, a dashboard and a few good standard charts probably can do the trick.
I think I linked to the paper on this project, out of Stanford and MIT,  a couple months ago. It’s a common runtime the creators say can boost performance on Spark and TensorFlow, among others. Here’s the project website and download info.
Listen to the Architecht Show podcast
Did you enjoy this issue?
The most interesting news, analysis, blog posts and research in cloud computing, artificial intelligence and software engineering. Delivered daily to your inbox. Curated by Derrick Harris. Check out the Architecht site at
Carefully curated by ARCHITECHT with Revue. If you were forwarded this newsletter and you like it, you can subscribe here. If you don't want these updates anymore, please unsubscribe here.