Oracle Goes Long on a PaaS – Now It’s Time For Some Blocking and Tackling Pierre Mitchell - October 6, 2014 2:23 AM | Categories: Cloud, Commentary, Technology | Tags: L1, Technology Recently I joined 60,000 others descending on this year’s Oracle OpenWorld 2014 conference in San Francisco and have been really immersing myself in all things Oracle, attending various sessions (e.g., rolling up the sleeves in some “hands-on labs” sessions) and talking to all sorts of customers, employees and partners. There’s so much to cover, but it’s best to start most strategically. It’s no secret that we strongly believe that the procurement providers of all forms will need to build their solutions on top of open cloud-based platforms that allow them, their partners and their customers to connect and interoperate without undue costs or headaches. At last year’s OpenWorld, Oracle hastily assembled a bunch of press releases surrounding the company’s cloud strategy, and most notably, its Platform as a Service (PaaS) offering. Larry Ellison was at a boat race (personally overseeing the Oracle/USA team take the America’s Cup) and missed his keynote – although he did apologize, and had even previously questioned the hype around the cloud when he asked “What the Hell is Cloud Computing?” But this year, he was on message, and on fire in classic Larry style. The message is all about the platform, the infrastructure to run it and the applications and ecosystem built on top of it. If I was playing the Bob Newhart drinking game, and “Bob” was replaced by “platform,” I would’ve had my stomach pumped nearly immediately. But the message, and the strategy, is dead on. To paraphrase, Oracle is the only provider with everything as a service that all works together: Infrastructure, Platform, Applications, Data, etc. Larry likes the term “Extreme Performance”, and it doesn’t just describe Larry’s keynote. There are a lot of really cool innovations in the infrastructure that include things like a columnar in-memory database (similar to SAP HANA), CPUs that protect data that sit in memory (e.g., this would have stopped the HeartBleed virus), hybrid storage (i.e., 4,000 GB in memory in a server node working in concert with Flash and traditional storage), and engineered hardware that can read 170B rows of data per second per CPU socket (that’s pretty friggin’ fast). All of these various platforms and infrastructures are of course “optimized” to work best together, and it’ll be interesting to see how standards-based and open/componentized the whole stack will be. Larry touted this openness while simultaneously crowing that 19 of top 20 SaaS players [can] run on Oracle infrastructure (i.e., Oracle DB or Java), and calling out Salesforce, SAP, etc., on this. And Larry of course had a good dig at SAP… "Ariba runs on Oracle. SuccessFactors runs on Oracle. They just bought Concur; it's moving to Oracle. I have no idea what runs on HANA, but it ain’t their cloud. That runs on Oracle. It’s rude, but it's the truth." Well, it’s sort of the truth for now, because SAP is expending a massive effort to get everything it can running on the SAP database. There were lots of laughs at these barbs, but also a bit of unease by some of the IT industry analysts in the room that I chatted with around a smiling bearded benevolent overlord touting the fact that everyone is running on his open infrastructure (note: “his open” is a bit of an oxymoron). As a side note, I was the only procurement analyst at the conference (no toothless bumbles here), a fact that I found sort of interesting. Even so, the poor analyst relations folks at Oracle are extremely overworked and understaffed – and much of it from the crushing weight of magic quadrants and the like – as Phil Fersht just wrote about here (check out the comments). OK, I digress. Back to Oracle. Interestingly, Oracle is basically touting the new SAP mantra of “Run Simple” in order to reduce costs, complexity, risk (e.g., security) and the dissatisfaction with traditional applications. Larry similarly did a mea culpa like Bill McDermott and admitted, “what has always not been part of the Oracle brand is ease of use or being low cost.” So, the focus is to “modernize” and move to the cloud the millions of business applications that on average are 20 years old and of which 75% are customized (statistics cited by co-CEO Mark Hurd). The big bet by Oracle is that it can create an “open” platform (database, app development tools, compute power, etc.) that it uses to develop its own applications while also selling the platform to the ecosystem of ISVs, systems integrators, BPO firms, and most importantly, to the corporate customers who can then more quickly use it themselves to develop custom/strategic applications that surround more commoditized vanilla functionality in the packaged applications. These applications (of which social, mobile, analytic, etc., are the big focus) can be from scratch or from a “modernization” process. Applying this PaaS model to procurement is something we’ve advocated for a long time (see related research links) because it allows firms to have an information architecture that can more easily plug and play various procurement providers at all levels of the “services stack.” This all sounds great and can actually be a major game changer because it changes the discussion from the promise of an ERP firm being a provisioner of a single integrated application suite to a single cloud computing provider that provisions a single platform and ecosystem upon which the application services, information services and business services can be assembled. But, there of course some issues here: Pushing legacy applications into the cloud doesn’t make them SaaS (i.e., ‘true’ multi-tenant architecture). A common PaaS doesn't alone solve this problem. Of course, the topic of what is “true multi-tenant” is a heated debate which we’ll return to in future research. When you do everything, you compete with everyone, so courting ISVs requires some clarity/finesse/trustworthiness. Buyers will want “loose coupling” across this stack and a true adherence to open standards so that the message is more than just “trust us – we’re Oracle.” Let’s address the first one. Larry went a bit off the rails and confused an aspirational strategy with reality when he said things like: “When you build your legacy applications from on-premise to the cloud, it becomes multi-tenant without changing a single line of code … We have so many ISVs who would love to see their applications be multi-tenant. Hey, move it onto 12c and your applications are multi-tenant. Move it onto 12c in our cloud, and they’re cloud multi-tenant applications … You have maybe a few levers you can push and you’re done.” Hmmm, as Inigo Montoya said in The Princess Bride, "You keep using that word. I do not think it means what you think it means.” I know the concepts of SaaS and PaaS are oft-debated, but I do believe that throwing your old applications and data (and data models) on highly scalable hardware (i.e., 'elastic' compute and storage) is really just a re-treaded ASP model circa 15 years ago. If his statement was true, then Oracle could just push that magical button and the flagship on-premise ERP suites that Oracle owns would now be multi-tenant! They’re not. And such cloudification of the applications is not so easy. Just ask Infor! Infor may even be out-platforming Oracle regarding co-existence/integration and modernization of legacy applications (e.g., UI design standardization, “wrapping” legacy apps beyond a Citrix-on-steroids’ approach, and even using OAGIS messaging standards for inter-application messaging no less!). In terms of issue No. 2, there’s a few important points here. First, I spent some time attending the Oracle PaaS sessions. The product is real, but still new. The PaaS is tied into the Oracle Cloud Marketplace where products can be developed, approved (by Oracle Product Managers to ensure that they are complementary), tested, certified and sold in the marketplace – and then deployed and managed by the developer (a “get app” feature is currently being developed that will allow Oracle to manage it like any other native SaaS application like the Oracle Fusion Procurement suite). There’s no reason to believe why this concept couldn’t be extended beyond applications to include complementary network services that could work with something like the Oracle Supplier Network. So, Oracle’s ‘business network’ could become an open ecosystem of interoperable network services rather than a re-badged eMarkeplace. But, this is where we end this piece where we started, which is about whether this PaaS is a “hail mary” or a new kind of “West Coast Offense” for software. We think the latter, and it has big implications for not just procurement applications, but for the procurement of business application services and broader business services. Still, the game is fought viciously on the field, and it requires commitment to move to an entirely new philosophy and new playbook. This will be no easy task for Oracle. If I were in the shoes of my clients, I’d be asking questions like: How will Oracle’s non-cloud procurement applications (and related applications in SCM, GRC/controls, PLM, project/portfolio management, cost management, etc.) REALLY get delivered in the cloud (beyond managed services partnerships like Enrich for Oracle E-Business Suite)? How will these applications make use of the PaaS – rather than just Oracle Fusion Procurement? Will Oracle Fusion Procurement be the new golden child app suite that will move upmarket, co-exist and then eventually subsume Oracle E-Business Suite? What partnerships will get brought into the fold such as those like Transcepta, Vinimaya, Prodigo, etc.? How will the Oracle Supplier Network get brought into the PaaS as a native integration service? What other areas of a “business network” have potential for inclusion in this PaaS? How will the PaaS help with application co-existence between the procurement product lines? Can/will it help? Will PaaS help improve the standardization on procurement information architecture areas such as portals, MDM, analytics, search, content management, dashboard/workbenches, roles, etc., that are currently managed “dynamically” within the siloed product organizations at Oracle? This is where an "Application PaaS" (aPaaS) managed by large SaaS suite player must translate to coordinated product development so that these areas are developed coherently across product lines (and even within them). What other PaaS/IaaS ecosystems are out there that can be tapped? Can Oracle really build a partner ecosystem in the cloud for Procurement / Supply Chain? What should Oracle users do based on this new Oracle strategy? How does the Oracle PaaS compare to SAP’s HANA Cloud Platform (beyond the database) in a procurement context? We don’t know all the answers to these questions, nor does Oracle itself, but we’ll try to address these issues and also some “blocking and tackling” analysis of the latest developments within Oracle E-Business Suite (EBS) and PeopleSoft Supplier Relationship Management. There are a lot of additional curious developments within the Supplier [Lifecycle] Management area and also with the Endeca extensions (within EBS… unfortunately). Stay tuned, but if you can’t wait, and want to chat about these issues, please don’t hesitate to reach out to me. Voices (2) Pierre Mitchell: 10.10.2014 at 10:33 am interesting development… http://recode.net/2014/10/09/google-and-snapchat-veteran-magnusson-to-lead-oracle-cloud/ Reply Daniel Ball: 08.10.2014 at 7:15 am Great article Pierre, never thought I’d see Inigo Montoya quoted on Spend Matters! Look forward to seeing some more comment on the (very valid) questions above Reply Discuss this: Cancel reply Your email address will not be published. Required fields are marked *Comment Name * Email * Website Notify me of follow-up comments by email. Notify me of new posts by email.