Whether it was a “mono-horrific architecture”, how “silos are for grain”, “cowboyism” development, or the frequent use of scatalogical emoji to reference it, the monolith took a beating today. The rigid stack may have been loosened by open source, the titles of those who administer it may have changed, but the monolith and its culture of protection persist. But the teams who run them are trying—and succeeding—to change it.
Day 2 of SpringOne Platform it was about taking risk, epiphanies, and building on breakthroughs. Event emcees Andrew Clay Shafer and Pieter Humphrey summed up the dogged commitment to that change in “we continuously deliver microservices or we die trying.”
And it was more than the technical sum of Spring, Cloud Foundry, microservices, and data services. It was about technology architectures and people systems that must be shaped and work in alignment.
How Pivotal Cloud Foundry, Spring, And Pivotal Labs Helps Developers Do What They Love
Onsi Fakhouri, VP of Cloud R&D, Pivotal
We’ll be putting the rest of the talks from SpringOne Platform on YouTube, subscribe to our channel to see them as they’re released.
Cloud Foundry engineering leader Onsi Fakhouri underscored that last theme, kicking things things off with a vision for software delivery. He told an engaging story about the potential of software—with its low cost of both change distribution—and the profound effects it can have on product teams. Some highlights:
- The human challenge of the monolith. A monolith is when developers add more stuff (their stuff!) to a single application or service and make changes complex. Operators put lots of machines together to make all that stuff work, and fight the fires that the machinery unleashes. So distribution gets complex, and teams try to avoid pain by shipping infrequently.
- Spring Boot: A framework that helps developers and operators deliver consistent, enterprise-class applications built atop an opinionated, productivity-oriented foundation.
- Pivotal Cloud Foundry: A promise to take any code, and run it, deploy it, scale it, and keep it secure.
- People are the final—and deciding—factor though. It’s more than the sum of PCF and Spring. Agile is a fighting style for sustainability—with minimal distractions, ownership of code, and high collaboration as the final factor for continuous iterative innovation.
Tech’s Rosie The Riveter Moment
Cornelia Davis, Senior Director of Technology, Pivotal
@cdavisafc sharing a pivotal moment in her life. #S1P pic.twitter.com/2VS7k6Yd6p
— ritam (@ritam) August 2, 2016
In the 1980s, when Cornelia Davis become a certified computer scientist she was joined by roughly 30% of her gender. Were she to come of age today intechnology, she’d be joined by closer to 18% of her female comrades. What happened? Cornelia offered up a spirited tale on the human smarts needed to take on the “other monolith” of assumptions:
- Assess your team’s implicit bias (it is a science today).
- Encourage young women to learn computer science and remove the stigma of being a female “techie.”
- Get involved!—with Girls Who Code or a similar program for young women who have the aptitude for this.
- Be an example in how you hire and tun teams. Both of these have a huge impact on a women’s initiative to get into and stay with computer science.
This is also a tale of two women: Cornelia, and her math-loving, still-in-school niece, who faces the assumptions that she and others are taking down, but holds the promise of what might be.
Microservices And Containers Can’t Fix A Broken Culture!
Bridget Kromhout, Principal Technologist, Pivotal
With a lifetime of experience in the operator’s’ world, Bridget Kromhout savers how things like containers are good tools. But they will not fix communication or team problems. A huge takeaway from this talk was the importance of decomposing the culture behind the monolith, not just the technical architecture itself. Highlights:
- You can deconstruct the monolith, but don’t forget the teams behind them.
- Microservices will not save a team dynamic: the communication needed on the technical implications and effects of these changes.
- Conway’s Law is still in effect: An organization’s systems will mirror the that organization’s communications systems (or patterns).
- “Why” you change is as important as “what” you change.
Customers
Comcast: Well Into The Dive
Greg Otto, Executive Director of Cloud Services, Comcast
Almost two years into his journey with a platform and Pivotal Cloud Foundry, Greg talked about their reality and lessons learned.
- Comcast lives PCF, with have tens of millions of transactions daily, and have business critical transactions on PCF.
- They kicked things off with a workshop (proof of concept) that lasted 2 days (build platform and deploy test app).
- Between 2015–2016, they spun up 4100 app instances and 900 apps.
- Comcast realized a 75% lift in velocity and time to market.
- Idea-to feature: weeks to days!
- Application scaling: months to minutes.
One quote among many: “For the first time ever, we fully upgraded cloud infrastructure with zero impact. In production. During business hours. Peak business hours.”
Citi: Establishing The Beachhead
Brad Miller, Global Head Digital & Cloud, Citi
“My name is Brad and I have a monolith problem.” And with that Brad from Citi gave an unvarnished appraisal of a monolith that blocks a creative and passionate team, and how his team is jumping into a big replatform shift. Highlights:
- Change can happen, and persist, because it’s backed by the leadership, the CEO.
- Catalogued familiar pain points: Many horizontal layers of proprietary infrastructure, testing consumes huge chunks of time, business as usual is unsustainable.
- Make surgical strikes: A small important team is starting to break through.
- Think like the customers and value them for insights into retails and other businesses.
- Bring the majority of work in-house, co-locate and concentrate teams.
- Bring compliance into the game to make Agile and DevOps work.
- Decompose slowly: Don’t build microservices that create redundancies (yes! And yes!).
- Building and training with Pivotal, simultaneously, is 3x faster than doing it in sequence.
Spring For The Ages
Phillip Webb, Spring Boot Lead, Pivotal
A fellow traveler gave a well attended walk through an era of change (in computing terms) on the microservices tool that has made millions of developers happier and more productive. It was another narrative on the power of feedback loops. The history:
- 2002: A response to the J2EE architecture cloaked in protection and risk aversion; Spring was (controversially) open source and gave the code & documentation away,
- 2007: All of the above comes online, complexity added to databases and servers. @configuration born, Spring 3; PostgreSQL and more OSS databases jump in; increasing more Linux OS usage.
- 2012: iPhones and Facebook change the landscape. Community feedback mechanisms on the website could have worked better. Spring is doing OK, “has high quality ingredients that we believe you can mix them yourselves,” but more bundling needed. Spring Boot was born. Big year for developer capabilities: Spring embedded Tomcat, Jar in Jar, Spring actuator for metrics—the beginning of coupling with larger ecosystem; Labs helps Spring create great developer site; Spring hosted on Cloud Foundry.
- 2013: Netflix OSS and Spring create a movement. It’s a catalyzing moment when a company adapted and ran on Spring, putting new meaning to surviving failure with an antifragile system.
- 2016: OSS has one has won, the cloud is complex. Spring is now another case for the feedback loops with hundreds of contributors, and thousands of pull requests.
The sample app showcased at the @spring1platform keynote this morning is available:https://t.co/wJoqgFgydP@springboot #S1P
— Stéphane Nicoll (@snicoll) August 3, 2016
More Thoughts In 140 Characters
Day one at #s1p has been really eye opening for this Domino Developer. There is a lot of great tech out there that we have missed out on.
— Declan Sciolla-Lynch (@qtzar) August 3, 2016
“Unlike most lists we make in software engineering, you need to pick all three of these.” @caseywest #s1p #springone pic.twitter.com/FWL2sfWXrb
— Bridget Kromhout (@bridgetkromhout) August 2, 2016
Among the many nuggets in @bridgetkromhout's #S1P talk: "put the why in commit messages, not just the what"
— Adrian Cole (@adrianfcole) August 2, 2016
"by simply dropping two pizzas into any organisation, teams will form around them, and then you have devops" – @phillip_webb #s1p @pivotal
— Amy Gdala (@monkchips) August 2, 2016
The scary thing about Spring One isn't learning about the cool things I'm not using, but the cool things I haven't even heard of yet. #S1P
— Gareth Davies (@wombleabroad) August 2, 2016
@caseywest Five stages of Cloud Native rocked! @spring1platform #cloudnative #S1P pic.twitter.com/XTpzONyxjc
— Melissa Chapman (@Bluesky142) August 2, 2016
It's rewarding knowing that you work for a company that's transforming how companies like Citi and Comcast do business. #s1p
— Craig Walls (@habuma) August 2, 2016
"Early on Docker said absolutely don't run this production. We were like oh well – YOLO!"@bridgetkromhout #S1P pic.twitter.com/Jk4bKomw3J
— Kenny Bastani (@kennybastani) August 2, 2016
Devops definition(s) @cote #springone pic.twitter.com/CcX4q6pcyq
— Barton George (@barton808) August 2, 2016
Recaps from other days
Read the rest of the recaps to see what else went down at SpringOne Platform: Day 1, Day 3, Day 4.
