Nice Little Pop - HDP

I can’t find any news on the 10% rise this morning, but I’ll take it.

Take care,

1 Like

I’ll take it.

Agreed. Only thing I saw is Zacks increased to “hold.”

Haha. But I’ll take it.


This morning the site issued:

HDP - Chad Bennett initiates on Hortonworks this am with a Buy rating & $23 price target. Chad believes HDP will prove to be a winner in what is largely a duopoly as they’re the true open-source provider for the next generation of IT services. There are some key tailwinds driving a robust 25%+ revenue growth outlook for the company… 1) the well-documented growth in data, but particularly streaming data where HDP’s new data flow product gives them a key competitive advantage over their competitor who offers no solution; 2) as a true open-source provider, customers who choose HDP can avoid vendor lock-in for data management solutions; and 3) HDP boasts a best in-class partner base that includes Microsoft (who is using HDP solutions at a large scale) and IBM (who has gone all-in on HDP and could drive $100M+ in annual revenues over the coming few years as the partnership reaches scale). Our $23 price target represents a modest discount to a peer group of other high-growth SaaS companies HDP init @ Craig Hallum … Content curated by … Have a question on this? Reach out to>

Got this link from Interactive Brokers.


Here’s a survey on big data which shows a tidy lead for Hadoop.…


2) as a true open-source provider, customers who choose HDP can avoid vendor lock-in for data management solutions

I don’t know enough about what HDP does and what their product looks like under the covers, but from my experience (Enterprise Architect at a Fortune 50 company) I would be surprised if this were absolutely true.

One of my jobs before I retired was to perform product reviews, make recommendations on enterprise IT s/w standards and vote on product standards. I also was a manager on a technology insertion project migrating the IT infrastructure from mainframes to distributed UNIX. On the technology insertion project my group established a lot of standards. All servers had to be POSIX compliant for example. And every vendor we considered claimed to be fully POSIX compliant - along with their own unique extensions. All DBMS had to be standardized SQL, and they were - along with their own unique extensions. The idea of setting the standards was to avoid vendor lock-in.

Good luck.

Try telling the DBA community that they can’t use imbedded procedures because they aren’t part of standard SQL. Those procedures forever marry the application to that specific DBMS product because a good deal of the programming ends up embedded in the DBMS, they become inseparable. Our initial migration used server from two separate vendors, but they were not interchangeable. Why? Because the network engineers used the UNIX extensions from each vendor.

Every vendor of an IT product advertises all the standards they comply with. They give the impression of portability. In fact, I’ve never seen it to be true.


The link Tamhas provided has me confused on multiple levels. I need to have a better understanding of the big data space in general, at least at a high level.

So maybe the following questions are easy for someone to answer, but the first chart in the link brings up a lot of questions. Here’s the link again:…

The link seems to mix and mingle companies and open source software. Some companies have more than one “big data source” such as Cloudera while others are lumped into one category like Hortonworks.

Hadoop is only used once “Hadoop Hive” and Apache is used several times. Those are all open source softwares and Hadoop and Apache are the same thing as near as I can tell (take this with a grain of salt as I could be wrong). So, these could all be a part of IT staffs using them at the company level. I dunno.

Also, the article mentions Hortonworks (9%) as a leader along with Hadoop Hive (25%). Cloudera seems to have two solutions each at 9% (total 18%). Again, this is all pretty darn confusing. Why wasn’t Cloudera mentioned as a leader?

If there is anyone that can shed any light, it would be appreciated.

Take care,


Well, A.J., with open source the company one gets it from is separate from the technology. Multiple people supply Hadoop because it is open source, one of which is Hortonworks. Apache is a source, a non-profit corporation ( ). Apache has quite a few different products, of which Hadoop is only one.

Does that help?

Does that help?

A little, but I framed my questions poorly so I shouldn’t have expected any great response.

First, I do understand the concept of open source. Anyone can use it and that is what Apache is all about.

Apache has quite a few different products, of which Hadoop is only one.

To be clear and this will help define the chart, Hadoop Hive is part of Hadoop which is part of Apache. So, Hadoop Hive should have really been called Apache Hadoop Hive on the chart to be consistent with all of the other Apache systems. The chart in this regard is very confusing especially when you consider that Hortonworks is employing these same products.

I’ll also note, just plain old Hadoop is not listed on the chart.

The following are all part of Apache:

Hadoop (0%), Hadoop Hive (25%), Spark (13%) (they didn’t put Apache in front of that one), Solr (9%), Sqoop (7%), Storm (6%), Drill (3%) & Phoenix (3%). That is 8 Apache open source products. And for those counting, they make up a whopping 66% of the overall “big data source” market per the article.

The remaining 44% of “big data sources” are offered by for profit enterprises some of which I haven’t heard of, but most I have. Again, based on this data, I’m surprised that Hortonworks was singled out.

Here’s the list of the for profit firms:

Clouder CDH (9%), Clouder Impala (9%), Hortonworks (9%), Oracle BDA (9%), Amazon EMR (8%)…etc.
I also just realized the chart items add up to more than 100%.

It seems like Cloudera has 2x the market share of any other for profit provider. Amazon and Oracle seem to be knocking on the door as well, but weren’t singled out.

At least I have a better feel for a very confusing chart now. I’ll still need to at least have some understanding of the different “big data sources” and what they do. For instance, I just learned a bit about Hadoop versus Spark. Spark, while it can run on other frameworks, works well on top of Hadoop. It isn’t necessary for all industries, but 10x - 20x faster than using Hadoop alone for certain needs. Obviously, this is an incredibly obtuse description of the differences, but it helps me understand to some degree.

Take care,

1 Like

Note that one of the big contrasts here is that one can use the open source version for free, but it implies providing all of the development and support for it internally. The only external “support” are open forums on the web.

The other potentially large difference is that the version from a for profit source is likely to have additional development added to it which may make the implementation and/or management of the installation significantly easier.

The reason ther are for profit companies in this space is because the open source software is difficult to employ without help.

I was surprised to see the level of open source software being used on its own.

The reason there are for profit companies in this space is because the open source software is difficult to employ without help.

Not really. I will give an example. Today, if you take a popular software like Oracle database there is so much skill set available you don’t need Oracle support to install, run (employ) the product (assume you own perpetual license) then why companies continue to pay support?

Support means if there is an issue (bug) or a change in laws, a security vulnerability, lastly an insurance policy for the IT department if something goes wrong. Of course, there are other things the produce vendor throws in.

In the case of Hadoop, the vendors try to add a few o their own tools on top of the open source to make it more useful and also they add hosting and some application support to make it a complete package for the customers.

Actually, many companies use a significant amount of open source software without support. Generally, support comes in when some critical business function runs on the open source.

1 Like

Kingran, I have to take exception to your characterization. For one thing, Oracle database is hardly open source, but rather is a very mature market leader in a domain that has become largely a commodity. So, of course, expertise is relatively available (to a degree, since I would also argue that most databases were poorly managed) and the product is very mature and stable. This is certainly not true of many open source products. Today, Linux may be getting there. I know that there are a fairly significant number of companies who will use one of the free distros of Linux because the OS is now quite mature and expertise widespread, but I also remember very clearly a time when my customers who considered Linux would only consider Red Hat because they wanted the assurance of support and a more mature package.

Hadoop is still pretty early in this cycle.

I have to take exception to your characterization.

Not clear to me what your exception is. I see your point about the product is not mature and stable therefore requires support is aligned and compliments the point I am trying to make. Am I coming across as something different?

Suffice it to say that I was trying to clarify that “it depends” rather than a blanket thing being true of all cases.