After attending a group discussion on vCenter I left perplexed by what was shared of the product roadmap and current customer feedback. On reflection I wondered whether something big enough was on the horizon that they were deliberately throwing us off the scent as I couldn’t believe how some questions raised in the discussion were not recognised as problems. Virtualisation management is so incredibly important, customer experience is key. Although great steps are being taken to address some functionality gaps, vCenter design-wise remains the same which isn’t good enough any more. I’ll go through the discussion and then add my thoughts.
This post adds some more colour to the highlights from my VMworld Europe 2015 coverage:
Since VMworld there’s been a huge amount of news about EMC(Dell)+Virtustream_vCloud Air. Basically EMC(Dell) will be consolidating most of its cloud assets under Virtustream which will be spun out as a separate company with EMC(Dell) owning 50% and VMware the other half. I’ll go into more details of the spin out but first here’s what I learnt speaking to them at the show.
Adding some more colour to the highlights from my VMworld Europe 2015 coverage:
I had a chat to Andy Sugden from Virtustream who EMC acquired recently. I had some misconceptions about Virtustream and how they competed with vCloud Air. Virtustream basically has software called xStream which big enterprises (on or off-prem) or service providers use to create managed instances of SAP and some other big applications. This takes a lot of the complexity away from managing these apps and provides a secure wrapper around the whole app up and down the entire stack so it can work in a multi-tenant environment and ticks all the security and compliance boxes. You can also create and manage performance and availability SLAs using the software. IBM use it (mmm, not for much longer I guess!). I can certainly see integration with vCloud Air by bringing some more of that secure multi-tenancy to vCloud Air but the big apps Virtustream manage aren’t the first applications people are moving to the public cloud so its normally a managed cloud offering, it was certainly interesting to get a better understanding.
I made a point as I always do at VMworld to head to the VMware Office of the CTO booth to have a glimpse into the future. I spoke to Jorgen Hansen about vRDMA.
This is an interesting research project within VMware to create a new class of distributed application in a virtual environment by allowing very fast transport by bypassing much of the VMware kernel and accessing memory on another host. This will allow applications to reserve VM memory via the hypervisor but be extremely scalable and fast, think HPC and financial trading. Expand the pic to look into the project:
Unifying Virtual and Physical Desktops with Synthetic Block Devices
Later on they also had a new research project to talk about: Unifying Virtual and Physical Desktops with Synthetic Block Devices. Rami Stern talked me through it which is all about having a single instance storage across the physical and virtual world so a single store of Mirage data as well as VMDK. Users would be able to move OS data from physical to virtual with very little data transfer, very much linking the different technologies VMware has acquired. Again VSAN is being looked at to do this, image deduped storage for OS + File + VM + Mirage + Cloud Volumes Data, again very interesting.
Presented by David Tennenhouse – Chief Research Officer (cool title), VMware
VMware has always been very academic research focused likely due to it being founding and having its head office in Palo Alto, home of Stanford University. Founders Mendel Rosenbaum and Diane Green were academics. VMware still maintains a serious academic programme partnering with researchers in systems and networks and invests significantly to maintain its innovation pipeline. David started the session talking about the challenges of being an industry leader where being a fast-follower is no longer an option. David went on to talk about some recent research trends and how he sees them shaping the industry.
More and more things are being virtualised including what he called heterogeneous computing such as GPUs and FPGAs.
There are new ways to build scalable and consistent distributed systems (he says this is having your cake and eating it)
David talked about the previous predictability for Moore’s law which allowed so many companies to plan their future on ever increasing compute resources. The increase of transistors is running out of room as the silicon space of the atoms is just getting too small. The only way to surpass this is to change how transistors are made. We will need to get improvements in other ways, in coordination across the industry. He said not to worry, it’s OK, there will be innovation beyond this with new materials.
I then spent some time at the fascinating VMware Video Game Container System. This is basically a demo of running MS-DOS in a linux container and then launching the awesome DOS Prince of Persia game within that container. Seriously impressive technology.
You can see the demo:
I spent time with Ben Corrie (the guy in the demo) who was the guy to worked it all out and put the demo together. He developed Project Bonneville which is running containers as VMs which is something I’ve spent a fare amount of time exploring at VMworld. Ben gave me a great overview of Bonneville and explained how docker and the container host and ESXi interact. I discovered that the stripped down VM appliance which is the base of the system is extremely light weight without even the docker client installed. As it is not on Linux but rather ESXi means you can run any VM that ESXi can run, hence being able to run Windows MS-DOS!
We chose MS-DOS 6.22 partly for nostalgia, and partly because it neatly encapsulates a simple legacy OS. In 48 hours, we were able to use a vanilla Docker client to pull a Lemmings image from a Docker repository and run it natively in a 32MB VM via a Docker run command. The image was built using a Dockerfile, layered on top of FAT16 scratch and DOS 6.22 OS images with TTY and graphics exported back to the client
This was held by Thomas Corfmat from VMware who currently works with vRealize CodeStream although the discussion wasn’t product specific.
Group discussions by nature aren’t presentations but rely on everyone giving input.
The session started with introductions from everyone to get an idea of what kinds of companies are in the session, as the introductions started quite a few people left immediately! Must be the pressure of a group discussion or wanting to stay anonymous!
Went through what was Continuous Integration / Continuous Delivery.
A poll was done as to where people were on the DevOps journey. 6% said they can release code into production within the hour, 60% were split between having no DevOps and starting to look at it and the rest were somewhere further along on the journey.
Some people shared stories about how they started, from changing from waterfall to agile approach, others being forced by the business.
The consensus was to start with continuous integration to get started, establishing toolsets like TFS or Jenkins for example and work on the people and process.
This session was delivered by Martijn Baecke and Robbie Jerrom, both from VMware, who were on the earlier panel. The session started talking about the transition from the client/server era into mobile/cloud. There have been huge changes in how software is being engineered, witness VMworld having a DevOps day today. There still needs to be a way to host both the apps of what they call yesterday (in reality 99% of apps running anywhere!) and the new generation of cloud native apps. VMware wants to ensure it can provide the platform to support both these worlds. The term used was future proofing so the idea is to build an infrastructure for your current workloads that can also be extended to deliver the new cloud native apps. The work needs to be done to transform current infrastructure but keeping one management later.
They looked at various strategies such as starting with a new cloud and rebuilding apps but this is a good choice for cloud native only, something not feasible for many companies. Greenfield can be done, starting again, but yet this is another environment to manage with upfront investment. Unsurprisingly they suggest using vSphere as the cornerstone of your infrastructure. Use software defined storage with hyperconverged/VVOLs and NSX for networking. Having a single platform makes it much easier to automate and orchestrate.
A great quote was given “Automation=Speed, Orchestration=Control”.
Guido Appenzeller – Chief Technology Strategy Officer of Networking & Security, VMware, Inc. Joe Baguley – CTO EMEA, VMware Paul Strong – VP & CTO Global field, VMware
Ray O’Farrell was sick so couldn’t make it, hope he gets well!
My notes, hard to encapsulate as it was a pretty broad discussion!
Paul Strong led the panel and went through each CTO and asked about their role
Joe: EVO:Rail, IoT and unikernels
Paul: Connecting R&D to partners and customers. integration back into R&D
Guido: Networking previously virtually integrated (same stack) unlike servers. Networking sales model changing to be like servers.
Local switching is easy, future is all about global connectivity with security built in.
Joe: talking about second effect, for example having virtualised networking, what will that mean, like cars being the enabler for Walmart which couldn’t exist without people being able to drive to out of town Walmarts yet the invention of the car couldn’t have predicted Walmart.
EMEA CTO Joe Baguley led the discussions with Martijn Baecke, Aaron Sweemer, Chris Sexsmith and Robbie Jerrom all from VMware which was to highlight VMware’s vision for next generation application development and hosting. The chat went through micro-services, 12 factor apps and how they could be deployed using PaaS and/or containers.
HA & FT are no longer needed. No resilience in infrastructure is required but pushed to the developers to create. This then becomes everyone’s responsibility to coordinate which isn’t an easy thing to do, I see more passing the buck!
There was some background on how agile makes things faster.
There was an interesting discussion for a typical use case for enterprise applications in that there’s no way you can make the entire application cloud native. There may be backend databases for example or even customer records on a mainframe which will never reach the cloud native world. Joe mentioned he is seeing a lot of traction for changing the middle of a 3 tier app so not messing with the back-end or the front end but breaking apart all the middleware into microservices to make them more efficient. Going cloud native for an application doesn’t have to mean going all in.