HDP News

I don’t know if you have experience in a Fortune 50 company which is in a heavy industry, but if so, I think you would agree with me.

Couple of examples: Few months back I was working with Oil Sands company in Alberta, basically capturing well information in Hadoop and that allowed the company to create new set of reports, which suddenly telling the company what wells will benefit from completion, workover, etc. I am not an Oil & Gas guy, but the company’s Cap-EX allocation has materially changed.

Working with an airlines company. You would be stunned to see the amount of data they generate daily and even worse they discard most of it because until now they had no way of either storing or making any sense of it. The company expects to save anywhere between $50 to $100M and a much better maintenance schedule, and thus reduce delays because now they can do lot more pro-active maintenance.

Modified IT system for a car rental company, the new system not only reduced the amount of time required for rental and return but overall saved more than $15 M on the first year of implementation the ROI is 100% on this project for the first year.

These are real time examples, where 100’s of millions of tangible benefit.

IT is a major differentiator.

Management is generally open to CRM, ERP, PDM, HRM, SCM,

All the above 3 examples are not yoru standard ERP, CRM. All these projects involve Big Data, Analytics, etc. The companies have automated those processes long time ago. They are moving into the next phase.

In the last earnings call IBM, mentioned how they have to rotate their workforce by letting go of 100K application support guys and re-hire same amount of workforce for digital.

Perhaps you are not in the industry now or you are not working with the so called “Fortune 50” companies.

1 Like

I would think of it this way, if an ordinary company doesn’t take in another order, they will have $0 million in revenue coming in. They won’t have that $155 million in revenue coming in… duh!

Hi Saul,

Yes, the revenue won’t come in per say, it already did and was deferred as a liability on the BS. Just because they have $155M in deferred revenue to recognized, it doesn’t mean they will receive any more cash. When a company is doing a lot of contracts bookings, they often collect the cash at the time of the booking and defer the revenue over the course of the contract.

Then I more interested in the booking trend and if it is increasing year on year then the revenue being recognized

That was my point, the cash flow will be king as you can see very different trends in revenue and cash when recognizing deferred revenue.

if an ordinary company doesn’t take in another order, they will have $0 million in revenue coming in. They won’t have that $155 million in revenue coming in… duh!

We had a similar conversation when discussing CRM’s deferred revenue.

The deferred revenue are either support (service) or license revenue. I don’t know why you think it is a big deal. Say for ex: a REIT enters a leasing contract to lease their building for 7 years, no one is counting all the 7 years of revenue on day one or for an Homebuilder a backlog of orders. The only thing that matters observation is the velocity of the deferred revenue. It should continue to grow or rather its grows sort of foretells the revenue growth.

1 Like

Better still is that many open source software can be used when working on Hadoop clusters, such as Hive or Pig (among many, many others and most are open source).

The products you have mentioned are part of hadoop eco-system. Hadoop and those products are all open-source products. The comparison with Linux is more apt for HDP. No enterprise IT department is going to buy an open source product because they need someone to go after for support. That is where HDP comes, they not only support their flavor of Hadoop, and enhance their position by offering additional tools, but they also provide implementation and on-going support. Lastly, they build industry specific solutions. These are primarily offshoot of the customer engagements they have. Say for ex: they work with an advertising company they gain some experience and realize the industry can benefit from a tailored or structured product, they create an industry solution.

If you look at Oracle ERP the suite only had core financial product and as oracle worked with many customers (and sometimes its partners) in different industries and build solutions for them, they started adding those features into their suite, is how it evolved to where it is today.

Will HDP become something like that? Needs to be seen. But there is a potential for that.

The big data industry has many players and there is a need for consolidation. So there is a bigger chance someone might scoop them before that. Given their market cap of $500M with close to $200M revenue run rate they are ripe to be acquired.

I never said it was a big deal, only that it might not be a key number too be excited about depending on when they collect payments

I have to agree with brittlerock that clueless beancounters can be a big problem for IT. Successful IT knows how to go above and around these beancounters to make their case.

My concern is that it should be clear that Hadoop is free (open source). Hortonwork sells consulting and add on software. That they are more nimble than IBM does not surprise me.

What Is Apache Hadoop?
http://hadoop.apache.org/

Denny Schlesinger

1 Like

brittlerock that clueless beancounters can be a big problem for IT. Successful IT knows how to go above and around these beancounters to make their case.

My concern is that it should be clear that Hadoop is free (open source).

Mine too. It is trumped by revenue growth and the anticipated growth in the data management market via Hadoop tech.

A.J.

Hortonwork sells consulting and add on software.

Actually, as I pointed out in this post (http://discussion.fool.com/hortonworks-32537042.aspx), Hortonworks does not sell add-on software, such as paid-for upgrades or connection modules. This is actually one of the dings against them right now, as those are common ways for companies to make money on Open Source projects.

If they do get into that business, it might be a nice stream of additional revenue and provide some lock-in dependencies to ensure continued renewals. But, there are no public plans for this that I could find.

I feel it’s pretty easy to walk away from support-only contracts: your existing applications keep working as is almost indefinitely - and you could even hire less skilled support people/companies to just keep the status quo going while you start something else for new or replacement applications.

But, if there were Hortonworks proprietary code in the loop, that code could require yearly licensing or even per usage charges (like Splunk). That gets harder to walk away from and so could increase customer stickiness as well.

2 Likes

I have to agree with brittlerock that clueless beancounters can be a big problem for IT.

The reality is the business directly goes to the market. They don’t even go to or through IT often.

Here is a $100 B company I am working with. They have 5 different major subsidies. While we are working with 3 of them, one decided to go with IBM and another one decided to go with Amazon.

So I was talking with their procurement guy about this, and he made an interesting insight that we are down selected not because IT preferred us rather business wanted us. If it was up to IT they would have gone with Amazon for all the divisions. BTW, we are no where near Amazon on price.

The point is neither IT nor finance had any say. Your descriptions are far from reality.

Hortonworks does not sell add-on software

HDP is the only pure hadoop player, i.e., they are the only 100% pure implementation of Hadoop. The next closest cloudera is only about 80% I would say.

In fact it is their significant differentiator.

one decided to go with IBM and another one decided to go with Amazon.

OK, that explains some of it anyway. I said I was an enterprise architect before I retired. The whole reason the great big aerospace company I worked for here in the Seattle area (that should be a clue about the “so-called” fortune 50 company I worked at) created an Enterprise Architecture group was to avoid the nonsense you describe.

Back when PCs just started entering business they were cheap enough so that they could be purchased by low-level managers. We soon had a dog from every town (more like 50 dogs). It was a support nightmare and cost far more than the simple cost of acquisition. Now scale that up to multiple content management systems, multiple MRP systems, multiple . . . You get the picture? It was now costing big bucks to host and maintain and interface all these disparate, expensive enterprise systems.

They changed the rules. No business unit could buy s/w or h/w without going through architecture. We had a process where we would put together a team of users, IT folks, Finance, Procurement and occasionally others to perform evaluations and reach a consensus on which to buy. I was focused on life-cycle management of information (and it’s digital doppelganger, data). But we also had teams for desktop, network, etc.

The whole reason for the existence of architecture was to reduce the overall cost of IT by reducing redundancies, planning transitions, looking into the foggy future, etc.

BTW, the reason I’m coy about my former employer is due to my role in the company, I was privy to a lot of proprietary stuff. I signed more NDAs than I can remember. I am very careful about how much I reveal and what I don’t reveal because even though I am retired, I am still bound by those NDAs.

4 Likes

that should be a clue about the “so-called” fortune 50 company I worked at

Yes. I was at the next door at Fluke.

My concern is that it should be clear that Hadoop is free (open source).

Since hadoop is open source organizations. A worldwide pool of enthusiasts contribute code to this. However, select organizations, people are allowed to write to the open source code and they are called committers. The below URL has the link of who those folks are.

https://hadoop.apache.org/who.html

What is important is look how many folks from Hortonworks are there. out of the 112, Hortonworks has 34. That’s pretty dominating position. This allows hortonworks a great influence on the release in terms of setting the agenda to deciding what goes in the final version to all sorts of mundane things.

Also, what this implicitly means is HDP will always have great knowledge and resource pool on Hadoop.

So this is not something I would be worried.

3 Likes

What is important is look how many folks from Hortonworks are there. out of the 112, Hortonworks has 34. That’s pretty dominating position.

Thanks CM. That helps. I noticed this last night when looking around their website and ran into the bios of the founders of the company who I recall were all committers.

I also recall one or two of them working on the “next” Hadoop which I believe was called YARN.

I get the sense they have an enviable talent position in this field and are prepared to continue innovating with open source tech.

Open source is a bit scary to me, I’ll admit. However, it may also be the new normal going forward. As with all tech, I’m sure Hadoop will be improved upon and that will likely come from the open source community again. If HDP is involved in that, they can continue to stay on the cutting edge and offer the best solutions for their clients.

Their roots coming from the open source community and staying keyed in to the open source community for data management tech, may give them a great leadership position.

Pretty much all speculation above…

A.J.

1 Like

Back when PCs just started entering business they were cheap enough so that they could be purchased by low-level managers. We soon had a dog from every town (more like 50 dogs). It was a support nightmare and cost far more than the simple cost of acquisition.

This is how Windows locked out Mac.

Denny Schlesinger

Open source is a bit scary to me, I’ll admit. However, it may also be the new normal going forward.

Open source harks back to the original hackers (MIT model trains) who thought that information should be free.

The Tech Model Railroad Club (TMRC) is a student organization at the Massachusetts Institute of Technology (MIT). Historically it has been a wellspring of hacker culture.[1] Formed in 1946, its HO scale layout specializes in automated operation of model trains.

https://en.wikipedia.org/wiki/Tech_Model_Railroad_Club

It’s also a backlash against the high cost of proprietary IP, specially software. It’s here to stay! And it’s one way that markets do their job of eliminating unnecessary profits. IBM bought into Linux, Oracle bought MySQL.

From an investor’s point of view it’s a game changer. With the availability of open source it’s much more difficult to become a “gorilla” because you don’t have a legal lock on the product. Gorilladom now comes from getting the network effect to work in your favor.

From the bits and pieces of information coming out of these HDP threads it’s becoming clear that Hortonworks has a good head-start in Hadoop but one should not expect returns like from the gorillas of old.

Denny Schlesinger

1 Like

I don’t know if anyone is still reading this thread, but one more comment.

Open source is not free of cost. It’s “free software” but internal support cost to any organisation can run from no big deal to a whole lot of money.

Any technology as complex as a data manager, programming language, etc. requires support. There’s configuration management and change management costs. You don’t want every IT guy (or end user for that matter) downloading whatever version they find out on the internet. That’s an invitation to disaster. It poses huge security risks and risks rampant incompatibilities. The source and version must be controlled and standardized. Version rolls must be thoroughly tested to ensure they play nicely in the environment.

Training is required. A support team and help request process has to be created. And on and on . . .

This is the value proposition of HDP. We can do it for with Hadoop cheaper than you can do it for yourself. Which is about half the truth. The reason it’s only half true is that it doesn’t really eliminate the need for internal support. No large company is going to roll to the next version just because they buy a support service from HDP. They will still need to perform all the same internal testing. They will still look for security leaks. They still need internal support specialists and so forth. Just maybe, fewer of them. The company I worked for was very reluctant to use any open source anything.

And of course all the same internal support was required of purchased s/w used enterprise wide. Where I worked, our shop had primarily settled on a single relational database manager for distributed processing. When the vendor released a new version, the CEO (a colorful man) always hyped all the new features. Our shop generally looked at that announcement as the revelation of all the new bugs that will be introduced, but with the hope that several of the previously announce bugs, er features, would be repaired. There were version rolls that we either refused to install or at least delayed until some of the most egregious bugs were fixed.

If the life blood of a company is information (it is), you have to be careful about transfusions and heart transplants . . .

1 Like

Open source is not free of cost. It’s “free software” but internal support cost to any organisation can run from no big deal to a whole lot of money.

You are looking at it from the customer’s point of view. When I wrote “free” I was looking at it from an investor’s point of view. My point is that Hortonworks does NOT get revenue from Hadoop like M$ gets revenue from licensing Windows or Oracle gets revenue from licensing their database. This means one should not use M$ or Oracle as a model for Hortonworks.

Denny Schlesinger

1 Like

Denny,

IBM bought into Linux, Oracle bought MySQL.

Is it possible for a company to “buy into” an open source technology? Or were you using “buy in” in the colloquial sense?

Thanks,
A.J.

Is it possible for a company to “buy into” an open source technology? Or were you using “buy in” in the colloquial sense?

Both!

I use MySQL so it was of interest to me to know how the Oracle buyout would impact me. This is a case of a company “buying into” by buying a target company.

I don’t use UNIX/Linux so I didn’t follow IBM’s move into Linux beyond noting it. This is a case of “buying into” without specifying the method used.

You’ll notice I said “buy into” for IBM - Linux while I said “buy” for Oracle - MySQL.

Denny Schlesinger