The thing about AMZN/Mongo

So…throwing up my hands asking for help here. I’m at a loss for why Amazon would do this. It seems to me they aren’t just competing with Atlas by offering to host Mongo cheaper than Mongo does. That would seem egregious but not surprising. Just typical Amazon greed – trying to “Wal-mart” MDB on price. To me that would make far more sense than what Amazon has done here.

But it appears Amazon is trying to use what Mongo has created to spawn their own spin-off “emulator” version of the software (database). Can someone help me out? Why would they do this? Do they really think their database is or will be better than Mongo’s? It’s a Mongo knock-off, for crying out loud! I just can’t figure out why they would choose to compete with the super-popular (#1) database itself. Especially since the thing is open source! Does anyone understand what I’m missing here?

Thanks,
Bear

10 Likes

Especially since the thing is open source! Does anyone understand what I’m missing here?
Thanks,
Bear

Bear:

There are two explanations and there is likely some contribution of both.

  1. AWS needs lock-in to their cloud and MDB is providing a database that is cloud agnostic. AWS wants their customers to stay at AWS and create lock-in to do so. They also likely didn’t appreciate the new license announcement from 2 months ago pertaining to cloud providers of its open source database requiring licensing. It would surprise me if they could respond that quickly to such a recent announcement so I suspect they have been toying with this concept for some time and thus this likely wasn’t entirely reactionary.

  2. There have been some large data set and speed issues with Atlas that were just updated this past month as announced in their earnings call. Note that these speed and dataset elements are exactly what AWS says they do better than Mongo…whether that is still true now after the Mongo update…we do not know. But there were some customers unhappy with performance with large data sets so this is what AWS tried address…perhaps again an attempt at lock-in vs the other clouds that depend more on MDB (though they all would like lock-in for sure).

We know AWS does not have an acceptable multidocument ACID so from a pure technology perspective, MDB leads by vast margin and keep in mind this was no simple engineering feat…3 years fulltime engineers to pull it off for MDB.

But IMO, the most important issue to come from all this will be:

  1. Recognizing that not every customer has been served satisfactorily with Atlas and these weak areas will be attacked by competitors in coming years

  2. Does MDB rise to the occasion and out-engineer these threats and continually look for customer experience improvements for what ails those customers.

Note that MDG has been largely mum on this development…lets see how they address it head on.

31 Likes

There have been some large data set and speed issues with Atlas that were just updated this past month as announced in their earnings call. Note that these speed and dataset elements are exactly what AWS says they do better than Mongo…whether that is still true now after the Mongo update…we do not know.

Thanks Duma. Sounds like MDB might have already fixed the issue before Amazon came out with this. It’s amazing when you look at the downloads just how much of a leader Mongo is. Amazon swooping in and creating a better database seems ridiculous – techies can correct me if I’m overstating that.

Note that MDG has been largely mum on this development…lets see how they address it head on.

I assume you meant MDB. The CEO said: “Imitation is the sincerest form of flattery, so it’s not surprising that Amazon would try to capitalize on the popularity and momentum of MongoDB. However, developers are savvy enough to distinguish between the real thing and a poor imitation.”
https://www.cnbc.com/2019/01/09/mongodb-falls-after-amazon-l…

Bear

5 Likes

From a purely non-technical and business standpoint, my working thesis is that AMZN saw MDB as an opportunity to get more folks onto AMZN cloud, but there was a problem… “MDB is too complicated to figure out and use”.

AMZN decides, “We can build a solution that makes the front end easier for Developers, and then we have them hooked to us for our AWS needs while meeting their DB needs”. AMZN can then offer this solution to developers as part of a AWS package of solutions.

I further suspect - cannot confirm - that AMZN has figured out how to patch in their other DB properties (aka ACID capabilities, etc) and realized they can offer DBaaS through AWS for NoSQL solutions.

So now, Amazon can offer a solution that sells/sweetens AWS contracts while solving an issue with the “ease of use” that MongoDB developers complained about. So now that AMZN can offer a similar service as MDB, why wouldn’t they start competing? They make money two ways (DBaaS and AWS) while MDB only makes DBaaS.

I have NO basis for this reasoning, NO knowledge that this has occurred, and NO confidence this is the case. But in my “worst case scenario”, this is exactly what has happened, and what it ultimately means, is that Atlas’ investment thesis has competition for developers on the AWS service, but holds true on all other services.

Again, this is a WAG

5 Likes

It’s amazing when you look at the downloads just how much of a leader Mongo is. Amazon swooping in and creating a better database seems ridiculous – techies can correct me if I’m overstating that.

IMO, while MDB does have a large number of downloads, obviously these are mostly fremium users…distinctly and vastly lower number of actual paid users.

If you look at what the revenue generation is for all databases annually…total database is around $70 BILLION annually of which MDB does $220 Million.

They are mere spit in the wind at this juncture so anything can still happen.

But as either Denny or Tinker recently opined, the fact that AWS wants to copy them…that was quite an industry nod…quite a nod indeed!

But MDB must respond with finesse and engineering prowess…….they are getting closer to being anointed the next modern database…but not as yet.

6 Likes

AMZN decides, “We can build a solution that makes the front end easier for Developers, and then we have them hooked to us for our AWS needs while meeting their DB needs”.

But they didn’t just make the front end easier. They built a whole other database. So their customers aren’t really getting Mongo’s. They’re getting Amazon’s new one. The two will diverge from here.

You later asked “why wouldn’t [AMZN] start competing?” To me this is exactly why: this is not low hanging fruit here. Mongo’s FREE product was best in class. Amazon could offer their DBaaS anyway, right? Why re-invent the database??

Also, I guess DynamoDB is now defunct?? DocumentDB replaces it? (It must be better because it’s based on Mongo!)

Bear

If you look at what the revenue generation is for all databases annually…total database is around $70 BILLION annually of which MDB does $220 Million.

How much of that 70 billion is NoSql, though? I think MDB is the big leader there.

Bear

1 Like

My thinking is Amazon is doing exactly what any company with aspirations to control and monetize their entire ecosystem would do. Think IBM in the mainframe heyday, they weren’t going to allow some 3rd party to control a critical core capability like the database in their environment - they delivered DB2 (& IMS etc). Since IBM also controlled all the underpinnings, APIs etc on mainframe they could tie the database very tightly with the platform, and since they had a captive client base they had some measure of control over the outcome. Customers always buy platforms when they can and like integration.

However in the distributed world things evolved differently, there were a myriad of hardware vendors (infrastructure) and numerous competitors in the database field independent of those ecosystems (Oracle, Sybase, Informix, Ingres, etc). The best combination of marketing/sales/innovation ended up dominating the space (Oracle) and part of the appeal was a single solution that was hardware and O/S independent running across many environments.

I don’t have an answer on how this will play out, I’m not well versed on the technical merits of MongoDB. Will customers more highly value an independent, portable solution that can be written once and deployed in multiple cloud environments (Amazon, Azure, Google, IBM) or will they more highly value the convenience and integration of a database that is more tightly coupled to a single environment like Amazon AWS? Is the NoSQL technology fairly settled in it’s capabilities (that would favor Amazon) or does it still require massive amounts of innovation as it evolves (that would favor a more nimble hyper focused company like Mongo IMO).

Amazon wants to create stickiness to their platform, so they will build in proprietary features that make it less likely you will ever want to migrate your cloud workload. That includes a tight coupling to the underlying database. Think Apple ecosystem, the more products you use the less likely you will leave.

Sorry raising more questions than answers, this in my mind comes down to how brilliant are the MongoDB developers and leaders, will they continue to add functionality that the market desires, with better scale, performance, reliability and portability that the me too vendors can’t match or keep up with. I was in the object database industry many years ago and watched as all those innovative companies and technologies were usurped by the relational behemoths that added “good enough” object to relational mapping features that while not true object databases, kept their infrastructure and skills relevant. This is why tech is so tricky to invest in, it innovates faster than most people can comprehend.

I’m still in Mongo but cut my exposure a bit to a normal not oversized position just to be safe.

8 Likes

How much of that 70 billion is NoSql, though? I think MDB is the big leader there.

Bear

Its about $4-6 Billion annually…again compare that to a mere $220 Million from MDB.

  1. They do not have the lion’s share of NoSQL but likely growing.
  2. The AICD product opened up a massive potential database market that dwarfs NoSQL

But again Bear, this is obviously a much longerterm story when viewed from a market share perspctive…which is why I would expect more future attacks on any perceived weaknesses in the coming years.

Which is also why I am VERY interested how they respond to the AWS threat…hoping for class, finesse and engineering prowess.

4 Likes

You later asked “why wouldn’t [AMZN] start competing?” To me this is exactly why: this is not low hanging fruit here. Mongo’s FREE product was best in class. Amazon could offer their DBaaS anyway, right? Why re-invent the database??

Also, I guess DynamoDB is now defunct?? DocumentDB replaces it? (It must be better because it’s based on Mongo!)

No, my comment is that they took MDB’s free code (because it was better and as you point out - more popular) and built the DynamoDB as the ACID component and bingo - they have their own DBaaS which (in AMZN’s opinion) addresses the ease of use concerns folks had with MDB while also reaping the benefits of the brand marketing and attraction of developers (until they can build DocumentDB’s reputation)

If the TAM is truly $70B, then it is large enough for AMZN to think about. Similar to the other “land and expand” models we explore, I’m not sure why this is a stretch to imagine? AWS is simply expanding the value add to customers who are part of the service. Plain and simple.

Time will tell. The good new for investors is that Tuesday at Needham will be very very telling.

1 Like

I’ve been trying to find info on the spend in NoSQL but either it’s not large enough to warrant major studies, no one has done it yet, or it’s locked behind pay walls or secrets.

If anybody has a link on that, would be appreciated.

In absence of market share data, a site DB-Engines provides a ranking of DB “popularity” that uses some sort of wizardry combining search trends, job openings with the DB mentioned, professional profiles listing the DB as a skill, discussions in DB boards, relevance, etc.
Read the link to the methods from the rankings link if you’re curious.

The DB-Engines Ranking does not measure the number of installations of the systems, or their use within IT systems. It can be expected, that an increase of the popularity of a system as measured by the DB-Engines Ranking (e.g. in discussions or job offers) precedes a corresponding broad use of the system by a certain time factor. Because of this, the DB-Engines Ranking can act as an early indicator.

Sounds to me like they are a measure of a Database’s relevance in the market and momentum.

That’s their claim, here’s the link to both the overall database rankings and the document store database in particular.

https://db-engines.com/en/ranking

https://db-engines.com/en/ranking/document+store

MDB is small compared to the huge top 3 (Oracle, MySQL, and MS SQL) but is rising really fast in the overall. And is ranked #5 among ALL databases. The top 3 are declining sharply.

On the document store (MDB type) they are far and away the leader and expanding the lead. About 6x higher score than next closest AmazonDynamoDB.

Of other note is the rise of PostregeSQL. And Elasticsearch search engine database. ESTC dominates search databases and is pulling away from Splunk and the cloud titan search applications. Gonna have to get more on that ESTC. I was reading some ESTC material and was curious on a statement about their licensing. The freemium ESTC license specified a size of the Elastic Node and limited the RAM and Storage of the data stored on the Elastic node. I guess I didn’t consider what they do as Database management, but maybe it’s closer to that than I realized?

https://db-engines.com/en/ranking/search+engine

Anyways, some color to the “market”. Not measurements of sales and installed base, but a measure of “popularity”.

Darth

11 Likes

Mongo’s FREE product was best in class.

If it is free, then why is MDB complaining about Amazon not paying for it?

1 Like

This is the crux of it, isn’t it?

Say I’m a photographer. I take an amazing photo. I decide that anyone is welcome to use my photo for their own personal use (e.g. print it out and put it on their wall), for free.

If you decide to take my photo, edit it/photoshop it, and then sell it on, I’m happy for you to do that without charge, but my only request is for you to share your edit so everyone else can use it freely.

This is the problem. Mr Bezos has taken my photo, edited it, not allowed anyone else to see the edit so they could use it themselves, he’s made lots of copies of the edit, and hires them out to people or venues to use, and gets paid monthly for the rent of these copies. Bezos says he’s not selling them so he’s technically abiding by my original wishes. This kinda pisses me off because I mean, come on…goes against the spirit of my wishes surely!

11 Likes

But as either Denny or Tinker recently opined, the fact that AWS wants to copy them…that was quite an industry nod…quite a nod indeed!

Wasn’t me but I agree.

Talking about downloads, I’ve used MySQL for close to 20 years, I’ve downloaded it two or three times and paid MySQL ZERO - ZILCH - NADA pennies.

Denny Schlesinger

1 Like

‘Your margin is my opportunity.’ ~ Jeff Bezos

2 Likes

If it is free, then why is MDB complaining about Amazon not paying for it?

It’s NOT free, it’s Open Source which means a license spells out how third parties may use it. In Mongo’s case, some of the code you have to pay for. There is a “free” version and a “professional” version. This week I downloaded software that is free for individuals but paid for business. The owner of the Open Source Copyright decides who and how the software may be used.

Throwing around the word “free” is misleading. Only “public domain” is unrestricted free.

Mongo is not complaining about “about Amazon not paying for it.” They are complaining about Amazon violating the license agreement.

Denny Schlesinger

20 Likes

Let us say Atlas use costs $10/month. $5 for hosting in AWS and Mongo keeps $5 to itself. Amazon wants Mongo’s $5 as well.

Amazon rev is about $200B/y. Database market is big enuf. for them to be interested. Moreover, it offers AWS lock-in too. So, it is a win/win for them.

Based on my example earlier if it costs just $6/month to use documentdb as service some companies will go for it even if the actual database may be a little behind. In the yahoo article they claim that documentdb service has a number of adv. over Atlas too. Don’t know if everything has been fixed.

https://finance.yahoo.com/news/aws-announces-amazon-document…

Customers …often only take advantage of a fraction of the functionality the API offers. Customers also find it challenging to build performant, highly available applications on MongoDB that can quickly scale to multiple Terabytes (TBs) and hundreds of thousands of reads and writes-per-second because of the complexity that comes with setting up and managing MongoDB clusters. As a result, customers spend a lot of time and expense managing MongoDB clusters at scale, including dealing with the undifferentiated heavy lifting of securing, patching, and operating MongoDB…

Amazon DocumentDB reduces database I/O by writing only database changes to the storage layer, avoiding slow, inefficient, and expensive data replication across network links. Together with optimizations like advanced query processing, connection pooling, and optimized recovery and rebuild, Amazon DocumentDB achieves twice the throughput of currently available MongoDB solutions.

https://www.epicgames.com/fortnite/en-US/buy-now/battle-roya…

Here is Fortnite. $3 billion profit in 2018. You can play for free! You never need to pay for it, ever, ever, ever. Yet Fortnite somehow managed to make $3 billion in profit. Did they do it on volume alone of people playing for free?

No, obviously people are willing to pay for added functionality. Same with MongoDB. When it comes to MongoDB it does not matter if millions use it for free because it is the tens of thousands or hundred of thousands of enterprise and institutional users that matter, because they are the ones that pay.

40 million downloads at last check. Lets say they get 200,000 enterprise and institutional customers some day. Wonderful investment for us! That is .0005% of all current downloads that are the only downloads that matter to MongoDB. Mongo’s business is not making money off of millions or even tens of millions or even 100s of millions of people who download and use Mongo (they can all have it for free), it is that very tiny small fragment of users that MongoDB makes money off of. Currently they have more than 8,000 customers and rapidly growing.

It is therefore misplaced to look at how easy it is to use for free!!! It is those that have need for the added functionality (i.e., enterprises and institutions) that need the added functionality and services. They are also the ones that have the most run-time and size of data stores on Atlas (and thus the ones paying the most for Atlas).

The beauty of Atlas is it enables MongoDB to monetize everyone - but even here it will be enterprises and institutions that are material.

Tinker

17 Likes

If you sell something on Amazon through the Amazon warehouses, they know how many you sell and for what price. If you are an independent seller, Amazon also knows how many you sell and for what price, etc.

Amazon knows how to mine all this data.

When companies are using AWS, Amazon knows some similar things about what applications you are using and for how long, etc.
Why would anyone think that Amazon isn’t going to cherry pick successful SaaS businesses and what to take more and more of the profit?

Mike

5 Likes

With time, I thought I’d look to source material from AMZN on why they even felt the need to invest their resources here.

Keeping in mind how customer focused this company is, I thought I’d find the cracks that they were seeking to fill.

https://aws.amazon.com/documentdb/

Self-managing MongoDB databases is difficult, time-consuming, and expensive. With Amazon DocumentDB, you can set up, secure, and scale MongoDB-compatible databases in the cloud without worrying about manually setting up and securing database clusters, running cluster management software, configuring backups, and monitoring production workloads.

Okay. So they were hearing about the challenges of managing these databases…implying that companies did not like having to staff up the internal expertise. Trend: O&M reduction at companies and outsourcing of services. We get this, this is SaaS in a nutshell.

Ok - but which customers

Shopping sites, online publications, digital archives, point-of-sale terminals, and self-service kiosks rely on content and catalog management systems to serve their customers. These systems need fast and reliable access to user reviews, images, ratings, product information, comments, etc. With Amazon DocumentDB’s flexible document model, data types, and indexing, you can store and query content (e.g., user reviews and demo videos for shopping sites) and catalogs (e.g., inventory lists for point-of-sale terminals and financial trades for trading platforms) quickly and intuitively.

Looks to me like e-commerce and small merchants. I also thought “sounds like all of Square and Shopify’s target customers”

Then I thought, “well Mongo knew of this issue because Atlas is doing the exact same thing”

So I conclude that the metric to watch is Atlas versus Amazon
DocumentDB as DBaaS

Finally, how hard it is to jump off Mongo DB and onto DocumentDB?

https://aws.amazon.com/documentdb/features/

Amazon says it will migrate existing MongoDB in “minutes” and offer “6 months free trial”.

Amazon appears to have locked itself into a near term sales battle of converting exiting self managed MongoDB into a DBaaS solution while MDB will be trying to convert those accounts into Atlas solution.

In the long term, I suspect that the codes will diverge as MDB license kicks into future evolutions

Needham will be very interesting to hear the response.

Just a Fool
Long MDB, but lightened position

17 Likes