VMworld 2015 Roundup: Day 4

Wednesday is typically a very breakout session heavy day for VMworld. There are no General Sessions and it's the last day for things like the Hand-On Labs and Solutions Exchange. Because the conference party happens on this night, there also aren't a lot of extracurricular activities available either.

#SDDC5027 - VCDX Unwrapped

This was a panel session to ask current VCDX's questions about the program, what to expect at a defense, etc. Let's hit a few of the points and questions I managed to capture.

• Multi-contributor designs are acceptable to submit for defense, but you have to make sure that you know all of the design. Even the pieces you didn't contribute to as much.

• What do you do if a customer asks for specific features in an infrastructure design but those features may not meet business requirements?

  • If this is being done in a customer-VAR relationship, the VAR typically creates a set of templates for each of the different design scenarios they're likely to encounter.

  • If you've already signed off on the business requirements with the customer and they're insisting on adding features after-the-fact, you may find yourself in a position to either defer to the agreed requirements (typically based on terms of your contract) or you may be able to have additional discussions about it (for example, by having a PCR created).

• How much longer will the VCDX-DCV 5 be available to submit for defense?

  • Typically two sets of defenses are available for the N-1 version after the N version is released.

  • VCDX-DCV 5 defenses will likely be available at least until 2016Q2.

• You can use cloud or VDI designs to defend for a VCDX-DCV.

  • It's highly recommended not to remove components not directly related to DCV (i.e. View) as you might end up with a design that loses some of the requirements and constraints that contributed to remaining decisions.

• Remember that the VCDX defense panel is there to help you.

  • They want to to succeed but you have to succeed (they won't "help" you pass, they grade you).

  • The panel asks probing question to assess your capabilities, they're not trying to tear down your design maliciously.

• How often do VCDX panel members research and read detailed technical material?

  • Typically only as needed.

  • Duncan Epping pointed out that when he first started working at VMware he felt he needed to get up to the same technical level as other VMware employees, so he read for at least 2-3 hours a day and answered questions on VMTN in order to both challenge himself and help others.

• You don't have to use the VMware design template for your design as long as you cover everything.

  • If you're a VMware partner, it's important to note that the set templates are not complete.

  • It was recommended to consider using the VMware template as it's familiar and that would make it easier for the panel to read.

• If multiple design methodologies are used and blended together in a submitted design, should this be called out and explained with the design, or will that add confusion?

  • Definitely add it, just don't create an unwieldy volume (ex. 800+ pages). After all the panel still has to read everything you submit.

#SDDC4593 - Ask the vExpert Bloggers

The Converged Cowboy makes an appearance

Another panel, this time with vExperts who blog regularly. Rick Shearer (RS) moderated, with Scott Lowe (SL), Chris Wahl (CW), Chad Sakac (CS) and Duncan Epping (DE) on the panel.

• Q: What do you use for your home labs?

• A:

  • CS: Has three vSphere hosts that he rebuilds every Christmas to keep his hands-on experience up. He pointed out that he doesn't use those systems as much and instead uses Cloud options like Pivotal Labs and AWS.

  • CW: Time is spent more on testing applications and OS's on his home lab. Now that it's been designed and implemented he doesn't do as much work on the lab (i.e. vSphere) itself.

  • DE_:_ Started with four hosts at home. Now he has 16 servers in VMware's Cork IE data centre. He spends most of his lab time on Nimbus, VMware's internal private cloud.

  • SL: In his home lab he has 40Gbps networking to multiple hosts as he focuses a lot on networking.

  • CS: EMC has vLab for partners and customers and VMware makes their HOL available. The era of home labs may be waning.

  • RS: Check out Ravello Systems for running hosted vSphere in the cloud. A discount/offer is available to vExperts.

• Q: What's "next"?

• A:

  • CS: The cloud native strategy announced earlier in the week is very exciting and is what's next for VMware.

  • SL: The Photon Platform is really disruptive.

  • RS: The entire SDDC stack is exciting.

• Q: Having Microsoft on stage for a General Session was very interesting. What partner relationships will we see next for VMware?

• A: Joke that was overheard after the General Session: "Netxt year Oracle will be taking the stage to explain exactly how their licensing works."

  • CW: Start learning how to code now.

  • SL: Learn other tech such as MESOS. Orchestration and automation is where things are going, hopefully VMware heads more in that direction.

• A multi-panelist discussion was had about Cloud Native Applications.

  • The trick is how to code this apps and deliver them as CNAs.

  • CNA is not new; AWS has already validated it as a model.

• Q: How do you decide what to write on your blog?

• A:

  • CS: Write what's going on right at that moment.

  • DE: Non-technical posts have surprised in how much response they've gotten and how well they've been received. Posts on topics like public speaking and career development.

  • CW: Write about what you want to learn. Chris' personal rule of thumb: He writes what he's right about and crowd-sources what he's wrong about.

  • SL: Will write something being worked on at work or that he's personally interested in when he has a few moments. If he can't write about it in such a way as he can explain it to anyone, Scott takes that as a sign that he doesn't know it well enough and has to go back and revise.

• Q: How to you continue to better yourself?

• A:

  • CS: Run towards the fear with a pointy stick. Be yourself, be authentic.

  • DE: Do what makes you nervous. For example Duncan used to be very nervous about public speaking. Listen to your customers, they'll tell you what to focus on.

  • CW: Face your fear. If you're intimidated that someone else knows something better than you do, remember that at one point they didn't know anything about it and had to learn.

  • RS: If you in an organization where they won't let you push forward and improve, you're probably working for a company that's the wrong fit for you. Don't be afraid to find your place.

#SDDC44595 - Ask the Expert Industry Titans

Another panel discussion, this time with folks recognized as virtualization industry titans. Rick Shearer (RS) moderates, with Kit Colbert (KC), Chad Sakac (CS), Vaughn Stewart (VS), and Nicholas Weaver (NW).

• Q: Looking 5-10 years into the future, where do you see storage, compute and networking going?

• A:

  • NW: True automation using machine learning and deep learning tied to compute and storage.

  • CS: Machine learning is dependent on memory. It needs to be smoking fast so put it on a memory fabric.

  • VS: We need new and better constructs for storage. There seems to be a server and storage admin convergence happening today (we didn't realize the fully converged DC admin predicted a few years ago and likely won't). We will be turning the data center back into one large computer.

  • WV & CS: Highly automated and disaggregated systems.

• Q: There's always been a separation of the other operations teams from the network engineers. Will things like NSX bring virtualization administration to networking, or networking administration to virtualization?

• A:

  • CS: Storage and server administrators are starting to come together, but network admins have been lagging behind. As NSX get stronger and network functions virtualization (NFV) takes off, friction will be created. Traditional silos and SLA'd environments don't work anymore.

#INF5116 - A Practical Guide to DevOps for VMware Admins

The third panel attended today was moderated by Trevor Roberts Jr. (TR) and made up of Josh Atwell (JA), Chris Sexsmith (CS), Egle Siegler (ES) and Adam Leff (AL).

• Started by asking "What is DevOps?" There are a number of slightly different answers depending on who you ask, and you have to adopt the definition that works best for you and your company.

• Check out the DevOps for VMware Administrators book from VMware Press.

• Configuration management is declaratively stating what you want in your server.

• Not every system is an artisianally crafted web server.

• Smart configuration management should only make enough changes to target systems to bring them into compliance.

• Tools like Chef and Puppet give different teams a common language which is very powerful.

• PowerShell DSC and Puppet now have interoperability. Each can call the other.

• Puppet Razor can be used to deploy bare metal systems including vSphere ESXi.

• Log analysis is very powerful and will help understand and control your environment. Check out the ELK stack.

• Back end of the session was the "DevOps Petting Zoo" were a number of audience members were brought up on stage to execute demos.

  • Audience member set up an NGINX website using Chef.

  • Audience member deployed bare metal *nix system, simulated against VMware Fusion.

  • Audience member used Git to commit a new web page and deployed it to the site via Jenkins.

  • Audience member would have (they ran short on time) used VMware CodeStream. Check out the CodeStream lab at the VMware HOL.

Mingling

I attended the VMware Canada Reception later in the afternoon and then it was off to the VMworld Party. It was excellent. Thanks VMware!