I was at MongoDB World 2018 for the last couple of days. I’m a little giddy, I drank the koolaid.
They tripled their research budget last year, and I think it really paid off. I had a chance to hear their CEO, CFO, CTO, and various heads of departments, and had some one-on-one time with the CTO. I also had a chance to have beers with some leaders of their teams after the conference.
One caveat - this was primarily a technology conference so they only talked financials briefly during the keynote (revenue growth 70% yoy, customer growth 80% yoy). We spoke mostly about the company’s business vision, technology road map, and product achievements. Here are my big takeaways:
Version 4.0 was released at the keynote, and it contains multi-document ACID transactions. The transactions are declared in the same way they are in SQL databases - no weird workarounds or special cases.
This is a game changer - ACID Transactions provides data integrity guarantees, while maintaining the speed and flexibility of Mongo’s document model. Working with Json schemas, which were released in version 3.6, this enables data designers to make Mongo collections as flexible or rigid as they want, enabling rules to ensure data integrity in certain fields, while continuing to allow flexibility in other fields. Note: ACID transactions on sharded collections are not currently available, but are scheduled for version 4.2
They had huge emphasis on Mongo Atlas - it was a big focus of their presentation. And they had numerous clients talk about it, including Coinbase (they handled the crazy growth in crypto only because they were on Mongo), Traveller’s Insurance transformed their entire company going from 2 databases to 2000 in one year. Braze does over a billion transactions a day on Mongo Atlas.
The best story was by Snag, the top hourly freelancer company in the US. They’ve been using Mongo as their sole database for 7 years - but the free community version, paying nothing for it. When Atlas came out, they did the cost benefit analysis, and have been on Atlas for the last year. They found out that they can save huge money using Mongo Atlas, rather than using the free version of Mongo and hosting it on their own servers. This is exactly the monetization story around Atlas I have been looking to see, and a number of companies had similar experiences.
Many new features on Mongo Atlas were also released at the keynote. The first was Global clusters, which gives you an amazing amount of control and granularity over where your data lives within Atlas clusters. It enables you to specify specific Mongo or other cloud titan (AWS, Azure, GC) data centres around the world to contain your nodes. You can then combine this with a document field that identifies a user’s location, to ensure that their data is always stored in specific data centres, both for efficiency and compliance purposes - GDPR. It also allows non-blocking nodes, which allow super fast reads, and hidden nodes which can be used for reads without affecting writes. It also allows you to be cloud independent with Mongo.
Mongo is making big steps into data analysis and visualization. They have really beefed up their Mongo Compass product (tied really well into Atlas), which basically allows data analysts to transform and load data into various front ends such as Tableau, or excel. They also have connectors that make the data in Mongo look exactly like SQL data and use standard SQL connectors. So now Mongo natively works with every data analysis tool on the market, whether you are using ETL, ELT, cubes, HDFS, MapReduce, Spark, or Talend. It works with tools that were built 20 years ago just as well as it works with tools built today.
A new exciting data visualization tool, Mongo Charts, was also was released at the keynote. It makes it very easy to create dashboards with charts and other data visualizations, through a point and click interface without needing any coding. The CFO demoed it, and created a couple of great dashboards in minutes just using point and click, with no code at all. This is very slick and easy to use, and sitting right in the database, it is super fast. It runs in a separate engine away from the database, and it can use special hidden nodes that stay current with your data cluster, so it doesn’t affect your database performance in any way - but still uses real-time data. There is no other data visualization tool that can do that.
Now there is a caveat about their data visualization - since it is based specifically in the Mongo database, it is only useful for data that is in Mongo. Most enterprise companies don’t have their data in Mongo yet. So it becomes a very attractive sales pitch to sell Mongo - by switching to Mongo you can not only get rid of your Oracle fees, but your Tableau and Alteryx fees also. But it is probably not going to affect sales of these companies significantly until Mongo covers other databases (extending the Compass tool) or more companies have their data on Mongo.
For me, the most exciting thing I saw, beside ACID Transactions, was Mongo Stitch. They describe this as backend as a service, and it is like adding a webserver to the database, so you can now access your database in a lot of powerful ways directly from the web.
Basically Stitch allows you to code 2/3 of an application right in the database. It is not like stored procedures, for the developers out there - those were always clunky and ugly and ran in the database, not always very quickly. Stitch runs outside the database in a totally secure web layer, allowing you to do amazing development and even integration with partners with minimal coding - very slick.
Mongo also released into beta a mobile version of their database which runs on all Android and iOS devices. This is a fully functional Mongo database. Stitch also has Mobile Sync, which allows you to sync, at all different levels of granularity, a mobile database on one device with another Mongo database - either mobile or regular.
On a personal note, I have mentioned in a post or two that I have a personal web application that uses Mongo, that is currently hosted on IBM’s Compose Mongo hosting service. Ever since IBM bought Compose, the service has become awful, and I’m getting ready to switch to new hosting. Now my website won an award from Microsoft, and I’m getting 3 yrs free hosting of pretty much anything I want, so I was going to port my database into Microsoft Cosmos, which has a Mongo emulator - it can look to my application like it is a Mongo database.
Even though I thought Mongo Atlas was very cool, I would have to pay for it, while Microsoft Cosmos would be free for me for 3 years. But after what I saw what Mongo is doing with Atlas at this conference, I’ve changed my mind. I’m going to pay for Mongo Atlas rather than take the Azure offering for free.
Like I said, I drank the Koolaid.