Yesterday I had an interesting moment where the available data was completely the opposite of what I expected. I wanted to test my assumption that leading with the name of an open source project would capture more interest than company’s distributions of that project. My thesis was that Cloud Foundry for example would be a term people searched for more than IBM Bluemix. I was wildly wrong
Helpfully Google Trends allowed me to disambiguate “Pivotal” to some extent. The graph is interesting because the recent marketing efforts of Pivotal and notably IBM are paying dividends for both parties. Cloud Foundry shows a gentle growth curve, with Pivotal trending up nicely. The Bluemix graph however is much steeper. IBM has been engineering and marketing resources into Bluemix since January 2014 – it is a Ginni Rometti level strategic initiative and it seems to be working, if search volume is a leading indicator of potential adoption.
IBM financial results were disappointing again this week but in terms of leading indicators Bluemix is definitely moving in the right direction, with interest growing sharply. I know Pivotal is closing some significant deals right now, though i have a less insight into Bluemix’s sales motion. But in search volume it appears that Bluemix is set to overtake Pivotal in short order.
It’s beginning to look a bit like the 1990s app server market, with Pivotal playing the role of BEA, as two major players carve out leading positions in an emerging market, with Cloud Foundry playing the Apache web server role. The best packagers in tech waves win big, and the PaaS market is now apparently a thing.
While we’re at it, let’s check out another project. It turns out searches for Hadoop do indeed dwarf those of the distribution companies, which represents the pattern I expected to see with Cloud Foundry.
IBM, Pivotal and Cloudera are all clients.
Obviously basing analysis on Google Trends is subject to all kinds of caveats. But search volume does tell us something about intentions and opportunities.
David Shacochis says:
October 22, 2015 at 2:21 pm
I think one of the things that explains the differences between the two project-vs-distribution charts is that the Hadoop one has a much tighter correlation between distributions and projects. Hortonworks and Cloudera have a heavy, heavy focus of distributing Hadoop to the enterprise.
Pivotal does more than just Cloud Foundry, with their agile consulting unit (Pivotal Labs) and their big data suite (Greenplum, Gemfire, etc) not to mention all the non-CF Spring assets.
Bluemix seems to extend far beyond CF as well, becoming IBM’s omnibus brand for all things cloud-native. Bluemix is CF, but Bluemix is also Docker, and it is also OpenStack VMs. The broadening of that brand results in searches that would never exclusively apply to the Cloud Foundry project.
jgovernor says:
October 22, 2015 at 2:29 pm
thanks David. good points. i was only too aware of the bluntness of the instrument I was using. but i think people searching for greenplum would still be using that term, etc. meanwhile Bluemix is stretching well beyond its initial CF roots.
James Watters says:
October 22, 2015 at 4:32 pm
Don’t forget IBM itself has 400k employees all wondering what “Bluemix” is–and they are also doing NFL ads about it.
For Pivotal we tend to use PCF as an advanced app server (exactly to your point) for things like Spring Boot, which has incredible, organic (no NFL adds here or even really any marketing) traction:
https://www.google.com/trends/explore#q=%22spring%20boot%22%2C%20bluemix&cmpt=q&tz=Etc%2FGMT%2B7
jgovernor says:
October 22, 2015 at 5:32 pm
hey james thanks for stopping by. IBM does indeed have massive resources on its side, but y’all don’t do too bad. As you know, RedMonk believes Spring Boot is going to be a legitimately big deal – that graph is ridiculous, although Spring + Boot has some ambiguity.
James Watters says:
October 22, 2015 at 5:44 pm
James–I was actually saying that 400k of those search results at a baseline were IBM internal 🙂
The deeper problem IBM has in this fight is that they DON’T have Spring Boot. Spring Boot (btw that search result is in quotes so no chance of anything else compromising it) is built for the cloud unlike Websphere.
If you want to transform Java for a cloud era, just sticking with Websphere is really problematic as it has incredible deployment complexity built for the slower moving, hand tuned, pre-cloud era.
IBM has done a great thing in supporting Cloud Foundry, but I also think they bring a lot more legacy to the poker table and I don’t envy their job of sunsetting Websphere (biggest cash cow) as they fight Pivotal for future workload relevance.
Thanks for watching the space, I suspect the next few years will get very interesting between Oracle IBM and Pivotal competing for cloud java.
Erin Schnabel says:
October 30, 2015 at 2:57 pm
WebSphere Liberty *IS* built for the cloud, so I’m not sure what you’re on about there.
“incredible deployment complexity built for the slower moving, hand tuned, pre-cloud era” does not apply to Liberty in any way, and Liberty is WebSphere.
jgovernor says:
November 3, 2015 at 11:54 am
hey Erin thanks for stopping by. What would you expect James to say? 😉 For what its worth his statements are about WebSphere classic, and let’s face it, even WebSphere CTO Jerry Cuomo is clear about the complexity drawbacks of J2EE and the App Server wave. Have you got some pointers to Liberty cloud deployments and customers at scale? I’d be interested to learn more.
jgovernor says:
October 22, 2015 at 5:52 pm
james- i knew what you were saying… 🙂
my assumption is that IBM will have to cannibalise its own J2EE revenues… with Spring Boot as well. It’s Apache licensed so there is nothing to stop them using it. Spring had already won, this is just the next stage of that.
You are also right to include Oracle in the analysis. Late to cloud, but moving faster now, and they certainly understand account control.
James Watters says:
October 22, 2015 at 11:14 pm
I’ll just leave this right here as well (for Oracle):
http://stackoverflow.com/questions/33287495/netflix-oss-spring-cloud-on-weblogic
Mac says:
October 23, 2015 at 3:19 am
Your use of Google Trends is wrong. Pivotal Software without double quotes will also include phrases that uses the words pivotal and software in any order. For example, it will also include the search “a pivotal moment in the software industry”. Please use “Pivotal Software” and “Cloud Foundry” in your search. Probably, you should include the term Cloudfoundry too. The data you are showing is not accurate for making the conclusion. Here is some help docs on Google Trends,
https://support.google.com/trends/answer/4359582?hl=en
jgovernor says:
October 23, 2015 at 3:11 pm
hey mt46steve. i am all to aware of the limitations of Google search trends, but in this case, pivotal software is a thing. i fail to understand how you can say the dates are wrong to reach a conclusion however. it starts far earlier that it needs to do, but the relevant trends are only too clear. thanks.
Mike Olson says:
October 27, 2015 at 2:52 pm
When I run that three-term search with the terms in quotes, I get “not enough search volume to show graphs”.
Rahath says:
January 21, 2017 at 11:30 pm
If a smaller company like Pivotal is bought out by a bigger company e.g. HP, the customers will suffer a lot of pain of migrations over time. If from start, worked with a big company’s cutting-edge product like IBM’s – who does most under their control from cloud to mid-tiers to mainframes, the customer companies will be safe from any buy-out chance (hope IBM will not be bought by Oracle or vice-versa).
jgovernor says:
January 29, 2017 at 4:12 pm
@rhahath given Pivotal is owned by VMware, EMC and now Dell, I would say it essentially deserves big co status.