Disclaimer: I attended Cloud Field Day as a guest of Gestalt IT. Whilst Gestalt paid for my travel, accommodation and other expenses, I am under no obligation to write anything about the event and the opinions below are entirely my own.
I’ve just finished up with Cloud Field Day 3 which has been a fantastic and fascinating experience. I’ve learnt a lot from all of the companies that presented, and whilst I may write longer posts specifically on each of the companies, I wanted to summarise the event by highlighting one thing from each company that I personally found interesting. Here are my thoughts about the companies that presented on Day 1.
Morpheus aim to provide a single abstraction layer across cloud infrastructure and infrastructure tooling, encompassing things like AWS, Splunk, Nutanix, Kubernetes, Terraform etc.
A key thing I took away is that Morpheus are not constrained by the lowest common denominator of only abstracting common feature sets across multiple service providers. When they develop an abstraction for a service, Morpheus will add services common with other platforms as a common abstraction, but they will also add services that are only available from that platform as a custom abstraction. Morpheus plan to periodically review custom abstractions and if they become something that multiple providers offer, Morpheus will then also add that as a common abstraction.
You can find David Estes from Morpheus specifically discussing this topic at CFD3 here; https://youtu.be/TtDbtAToVtE?t=21m53s.
Morpheus – That One Thing
Morpheus have a strategy for avoiding the problem of lowest common denominator, whilst also providing as broader common abstraction as possible. Whether Morpheus can continue to keep up with the frantic pace of new cloud provider features, in a cost effective manner for their underlying business model, for me is a key component to their success going forward.
Delphix aim to overcome the often lengthy and tedious cycle of getting database backups and restores from production into development and testing systems. Delphix use standard methods such as log shipping to maintain an almost real time copy of your production data within their virtual appliance(s). They then act very much like a virtual storage appliance, creating snapshots of that data and providing it out as mount points for dev/test database systems.
Prior to CFD3 I was already pretty familiar with Delphix and what they did. Its an impressive product that solves a real problem. However, I was a little disappointed by what they presented at CFD3. I was hoping for much more of a cloud focus, expanding beyond just replicating from on prem into existing supported databases platforms within EC2 and Azure VMs.
I would like to have seen a focus on some or all of the below;
- Support for managed database services such as RDS (more than just the current Oracle RDS support)
- Support for NoSQL services such as MongoDB, Cassandra and DynamoDB
- Expanding support across major cloud providers such as Azure and GCP
- A strategy for tackling new data sources and usage such as Amazon Kinesis and machine learning
Interestingly, Delphix stated they are actively working on these things, they just don’t have anything available to show, or even really talk about. Colin Rand from Delphix discussed some of their thinking in these areas at CFD3, you can jump into that discussion here; https://youtu.be/YcxuSJI3UcA?t=226
Delphix – That One Thing
I guess, for me ‘That One Thing’ from Delphix is that they didn’t really have anything cloud centric that stood out. Delphix have a customer base centred around large enterprises, fixing very apparent problems with traditional on premises database systems. That type of enterprise and databases aren’t going away anytime soon, so Delphix can afford a little time to work out and implement more cloud native solutions. However, whilst its something they are clearly working on, currently its next to impossible to establish if they have chosen the right offerings and will made them available at the right time as they have nothing to really show yet. I’d like to see Delphix back at Cloud Field Day in the future with a something a little more cloud focused to show off.
Oracle presented on Oracle’s public cloud offering, Oracle Cloud Infrastructure (OCI).
Oracle are joining the public cloud market very late and with the competition leveraging huge economies of scale to continue to rapidly expand. So how are Oracle differentiating themselves? It mainly seems to be a pitch to be the best place for a lift and shift of existing Oracle on prem workloads, and as a place to run high performance workloads, including on bare metal. This includes GPU accelerated workloads such as 3D modeling – something they did a live demo of which was very smooth – you can find that here; https://youtu.be/20qLlpZnCBs?list=PLinuRwpnsHaeMJYd_LdJ4PMSzz_8GEAmz&t=441
Oracle’s thinking is to fully enable any workload to run in the cloud (more specifically, in OCI), and therefore enabling enterprises to go all in even if that includes high performance Oracle databases that need bare metal compute.
Oracle have enabled some impressive features in their cloud for their target of high performance systems. The ability to use layer 2 networking and get a performance SLA do differentiate them from the likes of AWS and Azure. However, that really doesn’t compel me to want to use them for cloud native workloads that already work just fine on other cloud providers. OCI would be in my toolbox for application owners that feel unable to use other hyperscale cloud providers for high performance Oracle databases, or something that truly requires bare metal grunt, but I’m not sure what else I would use them for.
Oracle – That One Thing
The One Thing for me about Oracle’s CFD3 presentation was that Ravello was missing, something I was looking forward to learning more about. Perhaps this was down to Oracle having only presented on Ravello a few weeks ago at the Ravello Bloggers Day and therefore deciding to focus this presentation on their other offerings around OCI. Ravello and HVX came up a couple of times but not in any great detail, which is a shame, as HVX seems to be Oracle’s secret sauce for doing some interesting things across their cloud infrastructure. It felt like HVX, and especially its network virtualisation capabilities, will become more and more integrated with the OCI stack to natively enable things like stretched layer 2 in a region. That might help differentiate OCI from the competition for more standard workloads, but at present OCI feels like it is addressing a small but important niche; that of being the best place in the cloud to run Oracle databases.