Oracle have such a big target on their back. Everything they do will be received so critically by most of the market that it becomes one of those paradoxes whereby they end up exceeding (low) expectations by some margin.
I recently attended the Oracle Cloud Day in Auckland to hear their spin on the cloud-eat-cloud world and why amongst the usual suspects, they think they are in fact the real Kaiser Soze.
In a previous blog, the topic of the relevance (and probable demise of most) of enterprise clouds was discussed. This really puts the System Integrators in the firing line (Verizon et. al.) more than the traditional vendors. However, we are seeing some if not unholy, then at least atheistic alliances form – like the IBM (SoftLayer) and VMware recent announcement.
So what is the essence of the Oracle Cloud pitch?
What I heard was:
- “We are going to be bigger than AWS”, which, being kind, I’ll simply put down to blind optimism.
- “It’s all about the hybrid cloud, because there are plenty of cases where you don’t want to house the crown jewels of your business off-premises, even if that off-site location is a bank vault” - and I think this messaging is very much in tune with their customer base, whom have entrusted Oracle to run their businesses.
- “Let us modularise some of our technologies into SaaS offerings which can be consumed in a standard way” – again, I think this is on the money. These are not the systems of differentiation in a business, so they are safe candidates to effectively outsource and accept some compromises that come with using a standardised service.
- “Integration as a cloud platform service will differentiate Oracle Cloud” – I think it very much will. Oracle point out that the majority of integrations are now not inside the enterprise, but rather cloud-to-cloud or cloud-to-ground. This is a winner in my books, and more on that later.
- “Oracle Public Cloud Machine will be the way to run IaaS and PaaS on-premises in the same consumption model as the public cloud” – so yes they meant it when they said it is all about the hybrid cloud. And so too do Microsoft with Azure Stack and IBM/VMware with bare metal cloud SDDC deployments (all of these are not yet available). Oracle Public Cloud Machine was announced at Oracle World last year and they said it would be available in the US soon and in Australia by mid-2016.
So rather than “unpacking” some of these points, let’s grab the bucket of Lego and tip it onto the carpet (we can suck a few errant pieces up with the vacuum cleaner next week).
Logically enough the components of the Oracle Cloud portfolio include IaaS, PaaS and SaaS. In addition they referred to Data as a Service (DaaS) as an additional category where you leverage data as a standalone asset and connect it to other data sources. They talked about paying for an outcome such as the effectiveness of a campaign you run. I suppose it could be argued that DaaS is a subset of PaaS, however it is worth a specific call out from Oracle since it plays to one of their core strengths – owning the data and using it to add value to the business.
In the area of IaaS, the offerings are unsurprisingly average, but it is hard to catch the leaders of AWS and Azure when they have such a jump on you in terms of market share and incumbency, respectively. Focussing on the positives, there are features like bare metal (BMaaS) SPARC cores for computing, which seems like an excellent alternative to replacing massive (and massively expensive) legacy on-premises platforms that you are probably itching to port across to x86 (if you could only figure out how). However, there does not seem to be the maturity of supporting IaaS services for things like load balancing, automatic scaling, logging and auditing. At least they were not discussed from what I heard. I suppose the plan for now is that you would deploy these services yourself onto IaaS with a BYO approach.
There were also some intimations that Oracle IaaS would provide everything you would ever need, so you can forsake all other clouds. Given the maturity of the leading cloud offerings, this narrow view of hybrid = on-prem + Oracle Cloud doesn’t seem realistic. HOWEVER, Oracle have just announced a key acquisition. Oracle are set to buy Ravello Systems who make a hypervisor, HVX, which is designed to nest inside other hypervisors. Essentially this could (if leveraged appropriately by Oracle) make it easier for customers to port workloads into Oracle IaaS with this cross-cloud compatibility layer.
This is explained further in a letter from the CEO:
Ravello will join in Oracle’s IaaS mission to allow customers to run any type of workload in the cloud, accelerating Oracle’s ability to help customers quickly and simply move complex applications to the cloud without costly and time-consuming application rewrites.
Coming back to this concept of Integration as a cloud platform service (within the PaaS portfolio), I reckon this one has legs. The argument goes; the value of cloud is in the higher value services beyond IaaS – i.e. the platform (PaaS) that lets you build new applications and the software (SaaS) that lets you consume an outcome. So with the rise and proliferation of SaaS applications in the market, this means that companies will gravitate towards best of breed solutions, irrespective of whether they are integrated into a wider and coherent portfolio.
As an aside, it is the emerging approaches of micro-services (or re-emerging approaches if you are one of the ‘SOA already does this’ advocates) that has taught us that all we need is an API, three chords and the truth – and the world is our oyster of interconnected systems. But integration is not necessarily a trivial exercise, and has the possibility to suck time, energy, focus and resources. So a platform service that allows you to leverage integration tools in a standard way, and a marketplace that surrounds this where people can sell (or give away) the pre-canned integrations they have developed, sounds like it will be tremendously useful.
And for a final thought on Oracle Cloud for now, let me leave you with this. Just as Azure is a safe port for organisations intertwined in the value Microsoft, so too is Oracle Cloud for organisations that have ‘bet their business on Oracle’.
Yes, Oracle sound slightly delusional at times when they tow the party line and say you won’t need another cloud apart from theirs. However, if you move past this and recognise that leveraging Oracle Cloud will in all likelihood mean a multi-cloud future for your organisation (with some others components right sourced to other cloud locations), then you can go forth and extract value from features like; managed cloud services for the advantages of moving away from infrastructure management (DRaaS or managed Exadata in the cloud) or trying new modules (visualisation, analytics) or enhancing existing SaaS (integration as a service).
Each organisation will need to make their own assessment as to whether they use Oracle Cloud in a hybrid fashion, if they use AWS or Azure to host their Oracle workloads (in a BYOL or managed license mechanism). Or quelle horreur, they port some or all of their DB instances off Oracle and onto less proprietary (aka cheaper) DB types (like Aurora, PostgreSQL or MariaDB) using tools like the AWS Database Migration Service that can map your schema, stored procedures, views, the lot onto alternative database types – then populate that schema into a target for you – and then migrate the data off your legacy system. Now that is a killer blow.