Upgrade to Pro — share decks privately, control downloads, hide ads and more …

Navigating today's open source landscape

Navigating today's open source landscape

This presentation describes some techniques I've used over the last 5 years to gather market intelligence at the intersection of open source and cloud. The first version of this work was presented at the Open Source Leadership Summit in Lake Tahoe and a follow-up one at SCALE (this is the SCALE version)

José Miguel Parrella

March 10, 2018
Tweet

More Decks by José Miguel Parrella

Other Decks in Technology

Transcript

  1. What else has changed? ▪ Pace ▪ Direction ▪ Focus

    & research areas ▪ Motivations ▪ Collaboration still at the heart of it ▪ Ecosystem ▪ Commercial actors ▪ Community & governance ▪ Business imperatives ▪ VC funding ▪ Licensing trends
  2. Source: Joseph Jacks (@asynchio) ▪ 28 “open source companies” ▪

    Average age: 12 years ▪ $6.1B in VC funding ▪ Potentially making $10.4B in revenue ▪ Collectively valuated at $80B
  3. How we keep up has changed ▪ IRC ▪ Mailing

    lists ▪ Conferences ▪ LWN ▪ Slashdot ▪ Linux Journal ▪ Analysts ▪ Planets/blogs ▪ Forums ▪ Books ▪ Social ▪ LinkedIn ▪ Twitter ▪ Aggregators ▪ Hacker News ▪ lobste.rs ▪ Post-modern analysts & media outlets ▪ The New Stack ▪ Reddit ▪ Slack & Telegram ▪ YouTube & podcasts
  4. IT Devs How we contribute and release open source OSPO

    How we use and integrate open source CIO Why we do open source to begin with
  5. 3-step approach for CIOs 1. Stay informed 2. Sponsor your

    community 3. Bring open into your value chain ▪ Your community ▪ Your vendors ▪ Your priorities
  6. Building community ▪ Community is more than hosting a meetup

    – it’s also: ▪ Feedback cycle ▪ Attitude (different from policy) ▪ Skin in the game (investments, etc.) ▪ Ultimate goal: everyone in your value chain knows you’re open to open
  7. Sources ▪ 41% of the finds were on social media

    (LinkedIn, Twitter, Reddit) ▪ I use Nuzzel for both my LinkedIn and Twitter accounts (looks like it’s now also available for Slack) ▪ Top 6 sources: ▪ The New Stack (26%) ▪ Lxer (21%) ▪ Hacker News (19%) ▪ LinuxToday (12%) ▪ The Register (11%) ▪ Lobste.rs (11%)
  8. Business problem to solve Existing “polyglot persistence” workarounds are slowing

    down roll out of new applications Ideal business outcome Cloud-native applications converge in one multi-model database What’s this thing about? Multi-model open source databases can potentially outperform the specialized incumbents (MySQL, MongoDB, etc.) Who’s competing for the lead? ArangoDB CrateDB Postgres OrientDB CosmosDB (service) SAP & Microsoft likely to play a role Which of your competitors is doing it today? Thomson Reuters Accenture Comcast Moment of truth Application rollout no longer tied by incumbent installed capacity Community and governance Postgres has the most mature governance SAP has acquired Callidus (Orient) Arango is a consulting spin-off out of Germany What questions do you ask your vendors to get more out of this? Cloud providers: do you plan to offer <> as a managed service? If so, will you offer connectivity to other data stores, ETLs, big data, etc.? Incumbent database providers: what’s your multi-model strategy? Are you working with a particular open source project? Integrators and consultants: do you have production experiences with multi- model open source databases? Are you contributing to them? What questions do you ask your community to get more out of this? Are you attending any conferences/meetups covering multi-model open source databases? Have you run a Postgres POC? For which use cases? Have you assessed language support for multi-model open source databases? Have you assessed the licensing/IP implications of each particular database? What are the community health metrics of your shortlist?
  9. Business problem to solve Deliver apps faster, patch faster and

    guarantee infrastructure serviceability Ideal business outcome Cloud-native applications converge in one multi-model database What’s this thing about? Post-modern package managers for Linux have the potential to disrupt IT operations, but current technology approach and community strategy are divergent Who’s competing for the lead? Snaps Flatpak AppImage Container images Red Hat Canonical Which of your competitors is doing it today? Only ISVs: Heroku Microsoft AWS Sea Machines IntelliJ Moment of truth Application rollout no longer tied by incumbent installed capacity Community and governance Linux Foundation has a potential role, but isn’t actively looking into this OCI has an image-spec working group What questions do you ask your vendors to get more out of this? Cloud providers: will you offer a repository for [foo] in your CDN? ISVs: do you plan to offer a [foo] artifact to distribute your software? If so, by when? Which system are you going to use and why? What do you expect the artifact size to be in comparison to other distribution methods? How will your product manage upgrades? What questions do you ask your community to get more out of this? Do we operate DEB/RPM repositories today? For what use cases? How many source and binary packages are we building? How do we distribute those packages? Have we evaluated snaps vs. flatpack? Would we be able to distribute those use cases via container images or other methods? Is our CI/CD ready to produce artifacts in those formats?
  10. Resources ▪ Project repository ▪ github.com/bureado/open-cio ▪ GitHub’s Open Source

    Guides ▪ opensource.guide ▪ Linux Foundation’s TODO Group Open Source Guides ▪ github.com/todogroup/guides ▪ The Open CIO Series continues ▪ Stay tuned for OSCON, Texas Linux Fest and more!