It is hard for me to believe that I wrote my post “I Want a New Platform” over seven years ago. At the time that post led to our investment in MongoDB, which was then called 10gen. The company was started by Eliot Horowitz and Dwight Merriman inside of AlleyCorp, the New York based incubator started by Dwight and Kevin Ryan (Business Insider and Gilt Groupe also came out of AlleyCorp). Kevin had previously been an investor and board member in my own incubator and had become a friend. When he saw my post he reached out and we wound up investing. Today, MongoDB is the leading database to have emerged out of the NoSQL movement and I am thrilled that the company has just added additional financing.

At the time we first invested not only was the company called 10gen but it was also offering something a bit different: an integrated platform that provided not just a datastore but also a code execution environment. If you go back and read my original post that is what I had envisioned. In fact you could not access the data store by itself without the code execution environment. It turned out though that’s not what developers wanted. They wanted to be able to use their favorite libraries and to configure their own servers. Google App Engine which was first released around the same time as 10gen offered a similar solution and also struggled with developer adoption. In the meantime Amazon Web Services, which offered infrastructure on demand and had launched about a year earlier, took off.

As we are starting 2015 I am revisiting the question of code execution. Why? First, Google App Engine has made great strides. For instance, Snapchat was built and scaled on App Engine. Second, companies such as Docker, Mesosphere and Hashicorp are leveraging a slew of new open source projects and building their own software to dramatically restructure the code execution environment in data centers and on the public cloud. Their goal (maybe over simplified) is to make something akin to Google’s own internal infrastructure – on which App Engine was built – available to anyone.

The big question for developers now (and consequently for investors) is what makes the most sense going forward. One thing I am pretty sure about: managing at the individual server level will finally start to go away. As always I thought this would happen faster, as you can tell from a talk that I gave at Web 2.0 Expo in 2008 (I have to find my slide deck from then and put it online). What will take its place though is less clear.

Do people really still want to run their own infrastructure even at a higher level of abstraction? If you are an existing enterprise or a startup that has really grown the answer is probably yes, but what if you are just getting going?

I am currently working on a blog post about where I think we are headed. I would love to hear from anyone who is either working on this as a software or service provider or has an opinion as a developer.

It is hard for me to believe that I wrote my post “I Want a New Platform” over seven years ago. At the time that post led to our investment in MongoDB, which was then called 10gen. The company was started by Eliot Horowitz and Dwight Merriman inside of AlleyCorp, the New York based incubator started by Dwight and Kevin Ryan (Business Insider and Gilt Groupe also came out of AlleyCorp). Kevin had previously been an investor and board member in my own incubator and had become a friend. When he saw my post he reached out and we wound up investing. Today, MongoDB is the leading database to have emerged out of the NoSQL movement and I am thrilled that the company has just added additional financing.

At the time we first invested not only was the company called 10gen but it was also offering something a bit different: an integrated platform that provided not just a datastore but also a code execution environment. If you go back and read my original post that is what I had envisioned. In fact you could not access the data store by itself without the code execution environment. It turned out though that’s not what developers wanted. They wanted to be able to use their favorite libraries and to configure their own servers. Google App Engine which was first released around the same time as 10gen offered a similar solution and also struggled with developer adoption. In the meantime Amazon Web Services, which offered infrastructure on demand and had launched about a year earlier, took off.

As we are starting 2015 I am revisiting the question of code execution. Why? First, Google App Engine has made great strides. For instance, Snapchat was built and scaled on App Engine. Second, companies such as Docker, Mesosphere and Hashicorp are leveraging a slew of new open source projects and building their own software to dramatically restructure the code execution environment in data centers and on the public cloud. Their goal (maybe over simplified) is to make something akin to Google’s own internal infrastructure – on which App Engine was built – available to anyone.

The big question for developers now (and consequently for investors) is what makes the most sense going forward. One thing I am pretty sure about: managing at the individual server level will finally start to go away. As always I thought this would happen faster, as you can tell from a talk that I gave at Web 2.0 Expo in 2008 (I have to find my slide deck from then and put it online). What will take its place though is less clear.

Do people really still want to run their own infrastructure even at a higher level of abstraction? If you are an existing enterprise or a startup that has really grown the answer is probably yes, but what if you are just getting going?

I am currently working on a blog post about where I think we are headed. I would love to hear from anyone who is either working on this as a software or service provider or has an opinion as a developer.