VMware vCloud Director 9.1 is out!

So the VMware Cloud Provider BU has dropped the next release of VMware vCloud Director – version 9.1.

Release notes are posted here, but I’d like to summarize some of the great additions to vCD. I’ll probably miss a few things, but the below is what’s very interesting and shows the power of vCloud Director as we expand the platform.

vCloud Director 9.1 Release Notes

vCloud Director 9.1 Download

New Features

  1. Continued HTML5 transition – while this is a multi-phased approach, the following have been accomplished in this release. Quite a bit has been accomplished on the tenant side. Next release will focus on finishing the Provider side of the vCD management.
    1. Client Integration Plugin (CIP) for upload management – yay!!! Ability to upload OVF/OVA. 
    2. Multi-Site Navigation Portal – check it out. Very clean looking. I can also provide the organization association through the portal. 
    3. Create VM or vApp – nice simple workflows 
  2. Standalone VMRC Availability – great addition rather than the previous console access, which was always a pain. 9.0 released the HTML5 VM Console and now we have the standalone VM Remote Console support in 9.1. Again, no need for the vSphere CIP anymore with the HTML5 portal. 
  3. vRealize Orchestrator Integration – in my opinion, great addition to the vCD platform. Now we can provide direct vRO integration to vCD to kick off workflows. This is all done through the Content/Service Library.
  4. Python SDK and vCD-CLI – embracing the automation community. The SDK supports automation with Python, and the CLI enables Providers and tenant operations to integrate services within vCD. All open-sourced. Check it out here: http://vmware.github.io/pyvcloud/
  5. Container Service Extension – vCD will now support lifecycle management of K8s clusters through the Container Service Extension (CSE). K8s cluster nodes will be treated the same way as VM’s. One platform for both VMs and containers. Will be documented on the GitHub page also: https://vmware.github.io/container-service-extension/ 
  6. Support for SR-IOV / NFV Requests – this is a big item for our NFV friends, especially to guarantee network resources for low-latency demanding workloads. To add to this, we also added support for Large Page VM’s, guaranteed VM latency sensitivity for specific VMs.
  7. FIPS Mode for NSX – FIPS was introduced in NSX 6.3.0, but now we have the ability to toggle this within the vCD UI on per edge gateway. Obviously, you must be running NSX 6.3x or later for this to work.

Topics of Interest

  1. Moving to Cassandra 3.x for metric data – any legacy upgrade using KairosDB has Cassandra 2.2.6 support. Be aware of this for new installations.
  2. End of Support for Oracle Database – 9.1 will be the last release to support Oracle databases. I don’t see Oracle that often, but be aware of this for future releases!
    1. I would also advise all of my providers to get used to Postgres as the database option. We are trying to simplify vCD further….hint here.
  3. End of Support for vCloud API 1.5 and 5.1 – if you are using the 1.5 or 5.1 API for any API calls, it will not work in 9.1. Ensure you are changing any code before upgrading to vCD 9.1. Moreover, any API versions earlier than 20.0 will be not supported in future releases, so plan accordingly.
  4. Note that the SP Admin HTML5 UI is still underway. You will still continue to use the Flex UI for everything except the vRO registration and content library creation.
  5. There will be a patch release for Usage Meter and vCloud Director Extender shortly to support this release. Please be aware of this before any upgrades.

Another solid release from our team. I look forward to seeing this in production at our Cloud Providers!


vCloud Availability for vCloud Director – Tenant Walkthrough Video

I recently created a vCloud Availability demonstration video and wanted to share this out with others.

vCloud Availability (vCAv) for vCloud Director is a very powerful solution that provides Disaster Recovery as a Service (DRaaS) that’s built on top of vCloud Director. What’s great about vCAv is the ability to protect, migrate, and failover workloads from a tenant environment just from vCenter.

vCAv utilizes vSphere Replication as its replication engine while our Cloud Provider Business Unit built the architecture around vCD to provide scalable multi-tenancy. Granular selection of VM’s is possible that correlate to a DR-enabled VDC.

For VMware Cloud Providers that are interested in further details, here are some good links to start on vCAv:

Well, here’s the video. Enjoy!


vSAN Specialist Exam 2VB-601 – Tips

This was on my radar from last year but did not have enough time to get this done. I sat and passed the vSAN Specialist Exam today and wanted to share my feedback and experiences on how I approached it.

Read the Exam Preparation Guide

I know this is stating the obvious, but the blueprint should be your compass on preparing for this test. It reviews the Exam Sections and what to know before taking this test. Moreover, there are 10 practice questions to gauge your level of knowledge. When I started, I tested myself and knew what areas I had to prepare further for.

Download the latest guide here.

Read Storage Hub!!!!

I cannot stress this enough – storagehub.vmware.com is a WEALTH of information.

I went through each section and read each pertinent area of focus. Moreover, the Storage Hub has some nice features. Did you know you can export each respective section to PDF or mark it offline?

Very nice options for travel. I read quite a bit on my iPad and highlighted things I did not know.

What’s amazing is the depth and completeness of our documentation. I learned SO much by reading through many of these detailed documents. This should be everyone’s single source of truth for all things vSAN. Our vSAN Technical Marketing and Engineering teams have done a great job of making the inner-workings of vSAN public.

Do the Hands-on Labs!

There are three great labs available:

  1. vSAN 6.6 – Getting Started
  2. vSAN 6.6 – Challenge Lab
  3. Storage Policy Based Management

They are very comprehensive and cover your typical tasks within vSAN deployments and designs.

Exam Experience

I haven’t taken any of the Specialist exams before, so this was my first experience. I thought the test was fair, and not easy but not very difficult either. If you know vSAN, you should be able to knock this out. Key factors for preparation:

  1. Ensure you know how to size an environment. This is important in real life, but you should be able to calculate raw and usable sizes on the back of a napkin.
  2. Know your FTT policies.
  3. Go build vSAN in your lab and play around with it. I think this is critical to the learning process and also passing this exam.
  4. Last of all, know how storage operates. Having a storage background *does* help here.

It’s 60 questions and you have 105 minutes to complete, which in my opinion is plenty of time. I think I went through all 60 questions in about 35 minutes or so and was able to go through again and review all of the questions.

Good luck and onward to my next challenge!


VMware Cloud Provider – vSAN Best Practices

I’ve been getting a few requests recently on the best practices for vSAN inside of the Cloud Provider Program (VCPP). I’d like to spend some time on how VSAN works and on what works well within the consumption program.

First things first, how is vSAN licensed inside of VCPP?

vSAN consumes points just like any other offering inside of VCPP. However, compared to our bundles, vSAN is measured on Used Capacity on disk.

There are two base options of vSAN inside of VCPP:

  1. vSAN Standard
  2. vSAN Advanced

There are add-on packages for additional vSAN functionality. Essentially, this is added to the base option to provide additional feature sets. Two add-ons are available:

  1. Enterprise Add-On for the vSAN Standard base option
  2. Enterprise Add-On for the vSAN Advanced base option

This is all summarized in the Product Usage Guide (PUG as I call it) on page 45:

To summarize:

  1. vSAN Standard is 0.04 points/GB
  2. If I use vSAN Standard with the Enterprise Add-On: 0.06 points/GB (.02 uplift)
  3. vSAN Advanced is 0.06 points/GB
  4. If I use vSAN Advanced with the Enterprise Add-On: 0.08 points/GB (.02 uplift)

Now a single VCPP is equivalent to a specific cost, that’s all subject to your agreement. Moreover, Cloud Providers with higher commitments get additional discounts. This is on top of the current vSAN Promotion going on right now too. However, you can see how this can be attractive when we talk about consumption capacity and modeling.

vSAN Automatic Detection inside of vCloud Usage Meter

One of the great things about Usage Meter is it provides automatic detection of vSAN and will report on monthly consumption. This can be broken down to a specific tenant or customer based on your Usage Meter rule set.

There’s no additional configuration necessary to setup vSAN for Usage Meter reporting. Detected vSAN clusters will appear in the Cluster History report (depicted below).

Moreover, there are four features that are automatically detected based on the Usage Meter detection:

  1. Deduplication
  2. Erasure Coding (use of RAID5/6)
  3. Stretched Cluster
  4. IOPS Limit inside of the vSAN Storage Policy

There are a few things that are critical to accurate vSAN usage inside of VCPP:

  1. Ensure you are using the appropriate vSAN license key within the VCPP model. If you intend to use vSAN Advanced, use a vSAN Advanced key. Here’s why –
    1. Usage Meter will detect what license version of vSAN is being used. If an Enterprise key is detected, and either Dedupe/Erasure Coding features are being used, vSAN Advanced + Enterprise Add-On will be charged.
    2. This is important since if you are only intending to utilize vSAN Advanced (which includes Deduplication and Erasure Coding), you’ll want to make sure you pay for the appropriate amount.
  2. Stretch Cluster and Deduplication is applied at the cluster level. Hence, all workloads that reside on that vSAN datastore will be charged accordingly, even if you have workloads that may be tied to a PFTT=0 policy.
  3. Moreover, even individual features like IOPS limit and Erasure Coding (all within the SPBM) are enabled on a per-VM basis but scoped at the cluster level. If one or more VM’s are using these features, the feature is considered to be enabled for the entire cluster.

The below table showcases how vSAN is detected and the importance of using the accurate license.

vSAN Best Practices in VCPP

Here are a few things I’ve been messaging to my Providers regarding vSAN consumption inside of VCPP.

  1. vSAN is based on actual capacity used on disk. This is inclusive of the Fault Tolerance method used along with VM Memory Swap! Therefore, a FTT=2/RAID1 is going to consume a significant amount more storage than a FTT=1/RAID5EC policy.
    1. The chart that depicts storage capacity required for each FT method:
  2. VMs used the default Storage Policy are thin-provisioned while the VM Memory Swap is thick provisioned. Keep in mind when discussing this with eager tenants that like to over-provision their storage. vSAN will only consume the storage that has been written to disk.
  3. As for the VM Memory Swap, the Storage Policy does not impact the default setting. By default, VM Memory Swap is striped (or FTT=1) to another host. This can be overridden on a host basis by using the advanced setting of “esxcfg-advcfg -s 1 /VSAN/SwapThickProvisionDisabled”
    1. Review Duncan’s blog article more here: http://www.yellow-bricks.com/2016/02/29/vsan-6-2-sparse-swap-what-is-it-good-for/
  4. Deduplication and Compression can HELP from a cost perspective. Here’s why – Usage Meter will only detect what’s written to disk from the vSAN Cluster view. In the below example, I’ve written 101.72GB but with these efficiencies, I’ve written only 47.83GB to disk. From the VCPP program, I’d be only charged for about 47GB in space.
    1. ProTip: This is not a broad recommendation to use Deduplication and Compression just to save costs. Keep in mind the efficiencies only apply per disk group. Know your workload and I/O profile. This also applies to FTT policies – know your workload!

VCPP Sample Sizing for vSAN

So let’s go through some sample sizing on what an actual VM would be charged on a monthly basis.

Sample VM with FTT=1/RAID1 Policy

  • Sample VM: 2 vCPU, 4GB vRAM, 500GB HDD
  • SPBM – Standard Policy, FTT=1/RAID-1
  • Size Calculation
    • 500GB * 2 (FTT=1/RAID1) = 1000GB
    • 4GB Memory * 2 (Always mirrored) = 8GB
    • 1000GB + 8GB = 1008GB in total storage required
  • VCPP Cost Calculation – let’s say we are using the vSAN Advanced license which is .06 points per GB.
    • 1008GB * .06 points = 60.48 points per month
    • Let’s say my cost for a point was $1USD, this workload would cost me about $60 dollars a month.

Now let’s calculate this with an FTT=1/RAID5EC Policy. Keep in mind, all design rules apply still: still need a minimum of 4 hosts, dedupe/comp requires All Flash, etc.

Sample VM with FTT=1/RAID5/EC Policy

  • Sample VM: 2 vCPU, 4GB vRAM, 500GB HDD
  • SPBM – Standard Policy, FTT=1/RAID-5 Erasure Coding
  • Size Calculation
    • 500GB * 1.33 (FTT=1/RAID5EC) = 665GB
    • 4GB Memory * 2 (Always mirrored) = 8GB
    • 665GB + 8GB = 673GB in total storage required
  • VCPP Cost Calculation – let’s say we are using the vSAN Advanced license which is .06 points per GB.
    • 673GB * .06 points = 40.38 points per month
    • Let’s say my cost for a point was $1USD, this workload would cost me about $40 dollars a month. This is about a 32% reduction in savings compared to a RAID-1 policy. 

Now let’s say I applied a 1:5:1 deduplication and compression ratio to this sample VM. This is a conservative estimate since our standard vSAN calculator utilizes a 2:1 ratio.

Sample VM with FTT=1/RAID5/EC Policy with Deduplication/Compression Ratio of 1.5:1

  • Sample VM: 2 vCPU, 4GB vRAM, 500GB HDD
  • SPBM – Standard Policy, FTT=1/RAID-5 Erasure Coding
  • Size Calculation
    • 500GB  / 1.5 (dedupe/comp ratio) *1.33 (FTT=1/RAID5EC) = 443GB
    • 4GB Memory * 2 (Always mirrored) = 8GB
    • 443GB + 8GB = 451GB in total storage required
  • VCPP Cost Calculation – let’s say we are using the vSAN Advanced license which is .06 points per GB.
    • 451GB * .06 points = 27.08 points per month
    • Let’s say my cost for a point was $1USD, this workload would cost me about $27 dollars a month. This is about a 55% reduction in savings compared to the original model and 33% savings compared to the RAID5EC policy. 


  1. As you can see, vSAN consumption can be extremely granular from a cost perspective and will vary based on the protection scheme.
  2. Using the efficiency capabilities of vSAN along with Erasure Coding does save capacity, but remember what your workload requires!
  3. Usage Meter auto detects usage and will roll it up in the monthly reports – no additional configuration needed.
  4. As for pricing, there is a current promotion going on for vSAN inside of the Cloud Provider Program – reach out to your aggregator or VMware Business Development Manager for further information. This makes the overall vSAN pricing even more attractive on a consumption basis.