TL; DR: Solid progress for Cloud Foundry, the focus on business outcomes rather than just technology positions it well for the enterprise.
The 2016 Cloud Foundry Summit was held in Santa Clara on May 23rd to 25th, and we had the opportunity to attend both the summit and an analyst morning with Cloud Foundry CEO Sam Ramji and a customer panel.
Overall the customer panel was positive about Cloud Foundry. Customers of both Pivotal and HPE joined us and talked through their experiences of using cloud foundry, and the journey that they are on. The key part to understand here is that the journey is very rarely about the technology for these companies, it is about changing the entire culture within these organisations, a subject we will return to later.
Sam was keen to point out the open source nature of Cloud Foundry by highlighting the number of contributors to the project, and the growth in membership of the Cloud Foundry foundation, with new members such as Allstate, Volkswagen and Hazelcast announced. We were told that the overall growth of the foundation is being fuelled almost exclusively by inbound interest, which is a positive sign if so.
Cloud Foundry does set a very high bar on contributions. People have mixed feelings about this, for the purists this indicates a high quality bar, and requires a significant commitment to the project. For the occasional dabbler there is no doubt that the contribution process is off putting. I have yet to reach my own conclusions on this approach, but given the all in nature of Cloud Foundry the contribution process makes quite a lot of sense.
We also got presented with a set of slides telling us the market capitalisations of companies that are using Cloud Foundry, and positioning Cloud Foundry as the basis of the digital transformation efforts within these organisations.
Unfortunately, I did not have the opportunity to talk with Sam and drill into this market cap message. Either way it is, in my opinion, a pretty meaningless metric to call out. It ignores all the other technology investments an enterprise may be making and assumes cloud foundry is taking precedence over everything else for all things digital in these companies.
This is simply not the case in any medium to large enterprise. Indeed, only one company we heard from during the summit, operating completely from a greenfield position, stated they were all in for their entire product development strategy with Cloud Foundry as the underlying platform. Every other company or organisation we talk too has, as we would expect, multiple strands to their technology strategy.
One area that we found disappointing was the minor increase in new case studies and keynotes being presented by non-vendors. In both this summit, and in Berlin previously, there have been only four new companies talking each time.
We are aware of many more companies using Cloud Foundry than those that are willing to talk publicly about what they are doing. Hearing more about the how and what of what they are doing with Cloud Foundry would be beneficial for the entire community, and is something we would love to see more of at the next summit.
Releveling the Conversation
This is the key message about #CloudFoundry, focus on the innovative business outcome, not the infrastructure https://t.co/t1RXMbyyF0
— Fintan Ryan (@fintanr) May 24, 2016
The overall messaging during the keynote sessions of the conference was focused around the business outcomes enterprises are trying to achieve, and how developers can increase their velocity, rather than a discussion about the underlying implementation. Given the enterprise focus of Cloud Foundry this is very welcome.
The underlying technology is maturing very well, and is very interesting from a purely engineering level – but moving the general conversation up to the platform level away from things such as orchestration and deployment is something that the conference organisers are to be commended on.
The cloud native pie is going to be huge, and is already growing incredibly fast. The opportunities for all the players at different levels are immense.
Internet of Things
From my point of view, a big take away from Cloud Foundry summit was the amount of Internet of Things (IoT) related content that we saw. There were a number of very interesting IoT sessions with some genuinely impressive use cases presented.
Not unsurprisingly given the amount of travel all of us at RedMonk do, I took a particular interest in a session presented by Sarah Cooper of M2Mi & Shyam Nath of GE Digital on using Cloud Foundry with IoT in Aviation.
At the start of the talk Sarah pointed out the margins that the aviation industry operates on (the top level averages out at about 4%), and I was reminded of Michael Porters wonderful explanation as to why airlines are such a low profit industry. It is an incredibly hard industry to make money in over the long term, particularly in North America.
One example which was very interesting to hear about was the idea of combining data on checked in luggage transfers with frequent flyer data to see if it is worth delaying a plane five minutes to ensure a specific class of passenger’s baggage makes it on-board vs paying several hundred dollars to have the bag catch up with them. The actual mechanics of reaching this kind of decision and presenting it in a clear manner to a decision maker are very, very complicated.
But it was the sheer breadth of legacy systems and interactions that had to be addressed was impressive, from plugins for Oracles frequent flyer systems, to luggage handling systems and so forth, and the relative speed with which each interaction could be addressed and iterated upon that provided a very nice use case for Cloud Foundry and IoT.
Hybrid Cloud & Finance
One of the most interesting panels we attended was a discussion on Hybrid Cloud & Finance. The short version is that the cross cloud portability offered by Cloud Foundry along with the productivity gains made for developers once they start using the platform make it a compelling offering. It is also worth noting that none of the participants in this panel were interested in using Cloud Foundry without vendor support.
There were also some interesting positions taken during this discussion, in particular the public v’s private cloud debate.
What struck me most was how attached to their data centres some of the panel participants were. The initial arguments put forward against public cloud came around security, which was addressed by other panellists, and then moved onto a capex/opex argument. This part of the discussion, especially the security aspect, was like stepping back six or seven years.
That being said data locality and regulatory compliance were highlighted as big issues, and, in terms of risk management, this still remains a live issue. This is, however, predominantly an educational issue at this point – all of the major cloud providers are working to ensure these issues are addressed.
The reality here is that the lines between on premise and public cloud are blurring, and the longer term trend is in one direction for many, but not all, applications. That is not to say, there are fundamental laws of physics when it comes to storage and having proximity to data that can be ignored.
Acknowledging What Lies Beneath
In our discussions with Cloud Foundry users, at the summit, and in our broader day to day work, one message that has repeatedly come through is a need for Cloud Foundry advocates to respect and understand what has gone before. The legacy technology that lies beneath.
Almost every Cloud Foundry user I spoke with pointed out that they have a lot of legacy, and it is not going away anytime soon. As one CTO said to me
“We have maybe one hundred login services, we want to create a shared service to deal with this which is beyond cloud foundry. We do not want to be told to create a new service.”
Now several of the more traditional vendors who are investing in Cloud Foundry are head on in acknowledging these issues, Bill Hilf of HPE repeatedly made the point during his keynote presentation that large enterprises can’t just jettison the technology that is already in place and work with it as they develop new applications and models.
In our past conversations with SAP they frequently talk about using Cloud Foundry and microservices in conjunction with older SAP investments in enterprises. IBM are keen to highlight the work they have been doing with customers on connecting older infrastructure via the Bluemix platform, especially bridging between on premises and cloud with their most recent on premises offering.
In a similar vein to the existing infrastructure discussion, we also heard a lot of commentary about the disconnect between the agile approach sometimes advocated and the need for business units to have release dates for specific functionality.
Of course this is an incredibly common problem for agile development projects in many companies, but the commentary from several people we spoke with was particularly pointed here, one going as far as to say that there was such a significant gap in expectations and understanding that it risked derailing all the good work that had been done thus far as they tried to work with more parts of the business.
They were also very pointed on what constitutes an MVP in a developer’s mind versus what a business owner expects can often be vastly different, and while this difference remains there will be a level of inertia and resistance, beyond even what Doug Safford of Allstate termed as “The Frozen Middle”, to change in organisations that will remain hard to overcome.
The frozen middle – the biggest block to change in any company #CloudFoundry @doug_safford pic.twitter.com/uGp2s6pCeB
— Fintan Ryan (@fintanr) May 24, 2016
This is probably the biggest challenge facing not just Cloud Foundry, but the entire cloud native movement. Evangelizing and changing the culture of entire companies and facilitating business units embracing digital is always about far more than the technology. Part of that evangelizing is understanding where others in an enterprise are coming from.
A Single Request
Key takeaway from #cloudfoundry keynote & analyst session – community needs to create a repo of referenceable metrics for business cases
— Fintan Ryan (@fintanr) May 24, 2016
A number of Cloud Foundry users, both during the analyst session and in separate conversations highlighted the need for reference-able metrics to justify business cases within their organisations. In general, these are metrics focused on business value at a business unit level, not on something out of general IT spend.
If the Cloud Foundry Foundation can help create a repository of such metrics it will go a long way towards accelerating adoption in companies where Cloud Foundry is being considered.
Disclaimers: Pivotal, IBM, HPE, SAP, GE and Microsoft are current RedMonk clients.