Today marked the launch of Obvious Ventures, a new venture fund led by Twitter and Medium cofounder Evan Williams. Its mission is to invest in “world-positive investing.” Sounds like a nice company! And certainly Williams’ startup pedigree alone warranted its coverage by a number of tech publications.
But while every story on the launch was basically the same lightly rewritten press release with the same photo, the Wall Street Journal’s version was unique in one very notable way: It cropped the one female founding member out of the photo.
The Journal has since restored Virginia Sapp, the company’s operations manager, to the photo. And credit the outlet for transparency — it added a note to the bottom of the page stating that an earlier version of the story had omitted Sapp.
Nevertheless, there’s no explanation for why she was removed. It was probably done by some photo editor, but how come? Upon browsing through stories in that vertical, they feature a variety of different photo sizes and shapes. Yes, it was probably an honest mistake, but that doesn’t mean there weren’t implicit biases informing that decision. Why not cut out the guy on the left, for example? He looks like a member of the Swedish Olympic swim team, not a startup dude right?
I’m kidding of course, but you get the idea. These biases exist and they’re often difficult to avoid. And so the more we keep them in mind, particularly in the tech world which is overwhelmingly male, the better.
David Holmes is Pando’s East Coast Editor. He is also the co-founder of Explainer Music, a production company specializing in journalistic music videos. His work has appeared at FastCompany.com, ProPublica, the Guardian, the Daily Dot, NewYorker.com, and Grist.
We’ve entered a new era of enterprise computing, according Vinod Khosla. Ask the Sun Microsystems co-founder and Khosla Ventures founding partner and he’ll explain the thing that separates today’s environment from decades past is that nearly every critical application and process runs across a distributed system, rather than on a single machine.
“All network management solutions are designed for a client services world,” Khosla tells me. “Almost every part of the old stack has been changed in pretty dramatic ways. Google designed new kind of stack for itself, but this isn’t something other enterprises can buy.”
Mesosphere, a company founded by system architects from modern Internet giants like Twitter and AirBnB, has set out to build a data center OS (DCOS) that will enable everyday enterprises to operate as efficiently as the most sophisticated technology organizations in the world. Today, Mesosphere announced the details of its first ever commercial product, which will be available for purchase in early 2015, as well as $ 36 million in Series B financing to accelerate its product development and go-to-market efforts. The lead investor? Khosla Ventures. The round includes participation from existing investors Andreessen Horowitz*, Fuel Capital, SV Angel*, and others.
“I met them at the end of their last round. I told them then, I’m sorry we met you too late and missed out this time, but we want to be your first call the next time you’re ready to raise,” Khosla says. “It worked out nicely.” Speaking of the size of the round, he adds later, “They didn’t want to worry about money for a while and wanted to be able to build a big team. I don’t think we spend more than 5 minutes on the topic. We were more concerned with, ‘Can this become a big company?’ And I don’t think there’s any question the answer is, yes.”
Built on the back of the Apache Mesos open source project – the Kernel of which has been hardened by Valley stalwarts like Twitter, Airbnb, Netflix, eBay, and PayPal – Mesosphere allows developers and data center operators to look beyond individual machines, be they physical or virtual, and instead build and deploy applications and services across entire datacenters. This means the era of emergency pages going out late at night for failed machines or applications. Instead, Mesosphere can automate the process of spinning up instances of any necessary applications and processes on a new, un-utilized machine – something which is typically in abundance in any modern data center, given their poor historical utilization rates. As for deploying new applications, Mesosphere DCOS can take what would previously be a several day- or week-long process and reduced it to minutes, the company claims.
“We’ve taken a novel approach that means you can now program against an entire data center like it was one big computer,” Mesosphere co-founder and CEO Florian Leibert says. “There’s no need to write networking code and all distributed apps are managed under the covers. One question we were asked frequently early on is, ‘Are we too early?’ We feel like the timing couldn’t be better. We’re perfectly intersecting several important trends. No apps run on a single machine today, everything is distributed system. And it requires lots of work to make that robust, and it gets harder as these systems get bigger. Doing this manually is really inefficient and not terribly fault tolerant. The average data center utilization rate today is between 8 to 12 percent, which is abysmal.”
“There’s a lot of things going on in the data center that people haven’t realized,” Khosla says. “For example, people didn’t realize that virtualization was important until it was too late. People today are still talking about discreet functions, not operating sytems for the data center.”
Mesosphere DCOS includes a command line and visual dashboard user interface, as well as key operating services systems like finder and storage. An accompanying API and SDK means that developers can build their own scalable applications to run on top of the system. Mesosphere natively supports distributed applications like Apache Spark, Apache Cassandra, and Google’s Kubernetes, as well as legacy Linux applications. Also, Mesosphere can run on premise on bare metal or in a virtualized private cloud environment, and works with third-party cloud platforms like Amazon AWS, Google GCE, Digital Ocean, Microsoft Azure, Rackspace, and VMware vCloud Air.
Mesosphere has raised $ 46.5 million to date and grown its team in both San Francisco and Hamburg, Germany to a total of 40 people. Leibert expects to see that number rise to 120 by the end of 2015, with growth coming across product, engineering, and sales.
A beta version Mesosphere DCOS is already in use among several unnamed “lighthouse” customers, according to Leibert, who says the company is seeing demand among Global 2000 corporations within the financial services, consumer electronics, and industrial manufacturing sectors, among others – all of which “share a need for a large numbers of data centers, and which view software as an area of differentiation,” he explains. The company is still ironing out its pricing structure, althought Liebert reveals that it will likely be based on acombination of usage and value delivered – meaning quantifiable cost savings as a result of deploying Mesosphere. The first version of Mesosphere DCOS is targeted at the enterprise, but going forward, the company plans to add a free, Community Edition product and will continue contributing to the open source community.
“We’re finding that companies are adopting Mesosphere around a single data set or innovative project that are particularly ripe for switchover,” Leibert says. “They start small, but as they see the value their cluster grows. We allow engineers to focus again on writing software rather than restarting software on machines that fail – the ROI is clear.”
[*Disclosure: Andreessen Horowitz partners Marc Andreessen, Jeff Jordan, and Chris Dixon are personal investors in Pando.]
An evolution of virtualization. Rather than provision and manage virtual machines, Mesosphere’s data center operating system makes executing and managing applications in a data center just like launching applications on a smartphone by running them on Apache Mesos. Portable across data centers. Mesosphere’s data center operating system runs on popular, public cloud providers as well as directly in private data centers. Easy to run and scale. Mesosphere’s data center OS makes it easy to run big data frameworks like Storm, Apache Spark, and Apache Hadoop. Applications designed for elasticity can leverage idle resources, improving resource utilization and reducing total costs.
Michael Carney is a West Coast Editor at PandoDaily, covering venture capital, financial technologies, ecommerce, on-demand services, and the future of television, among other subjects. He has spent his career exploring the world of early stage technology as an entrepreneur and early-stage investor, working in multiple countries within North and South America and Asia. He is an enthusiast of all things shiny and electronic and is inspired by those who build businesses and regularly tackle difficult problems. You can follow Michael on Twitter @mcarney.