Month: March 2023

Posted on

Cluster Templates / Expert Libraries

 

Cluster templates and Expert libraries: 

What is a cluster templates? 

Sizer 6.0 has come up with a concept of cluster profiles. These are specific cluster settings for a sizing.  The cluster profile is user-defined and users can create as many profiles as needed by giving the profiles a unique name.

How does creating a cluster template help ? 

Different types of workloads have different preferences for the recommended solution. For example – a Database cluster would be quite different from a VDI cluster. Whereas a DB cluster would almost always be All Flash or for some SEs even prefer All NVMe nodes, a hybrid node for VDI cluster would be sufficient in most cases. These are different options in the cluster settings that can be set for the given workload by the users.

Another example is preferred set of models. Some customers would like to go with certain models (already existing in their environment) or filter out certain models (for ex: 1065 as it is single SSD).

Currently, SEs change these filters and preferred options in the cluster settings page every time for a new scenario/sizing.

With Sizer 6.0, Nutanix SEs and partners can save these settings (as a cluster template) and use it for future sizings.

For example: There are certain settings a user typically prefer for Oracle

1. Nodes have to be All Flash (or NVMe)

2. Should be a hot processor (>3.0GHz)

3. Only certain models (NX-8050 , NX-8070)

SEs can save an Oracle cluster profile with these settings and import this profile every time they size for an Oracle workload. This helps avoid repeating the settings every time making the sizing process faster and consistent.

Power of Cluster template/Expert Library:

Leveraging Nutanix expertise: 

Why care?

 

  • Focus on the Workloads with the customer and then let your preferred cluster settings control Auto sizing and yield strong solution

 

  • Can get more consistency across wide range of Sizer users
  • You don’t have to know all 18 vendors in Sizer (focus on workloads and software and less on HW)
  • Can easily compare multiple vendors

Brings uniformity to solutioning

  • Across different user persona(expert, moderate, beginner)
  • Across internal SEs and partners
  • Removes the need to be Nutanix technology expert or expert of the workload being sized(eg Database)

How do I create a cluster template?

There are two ways to create a user-defined cluster profile in Sizer.

  • Directly creating cluster profile from user preferences
  • Saving an existing cluster setting as profile while doing a sizing

Steps for creating user defined templates:

  • Directly creating template from user preferences

Go to the User icon on the top right corner and click Templates (pic 1)

 

There are two approaches to templates.

Under ‘Templates’ , click on ‘ClusterTemplate’

Click on Create in the next screen (shown below)

 

Next, on the cluster settings screen, enter a name for the profile (eg Database cluster template) and select settings (certain CPUs, CVM allocations, preferred platforms and many more attributes shown in next section)

Click “Review” and then “Save” the profile which starts showing in the Cluster template dashboard (shown below)

  • Save Template while creating the scenario

To provide flexibility and for users to be ‘in context’ while creating/saving a template, additionally, an option in Sizer is to save a particular cluster setting done for a current sizing/opportunity as a template for applying in future for a similar opportunity or workload type.

This is provided as an option “Save Template” in the cluster settings page. It saves the current settings in the page (user needed to provide a name for the template)  as shown:

 

 

Cluster template attributes:

Below is the list of attributes and filters which can be set for the cluster profiles based on the requirement of the cluster/customer.

Cluster Profile attributes sample values
Data Center models / Robo Models NX 1065-G8/1075S/3060…
Minimum CPU Freq(in GHz) 1.7
CVM per Node
         Cores(physical) 4
         RAM 28
Cluster type Homogeneous
Failover Standard(N+1)
Storage Any
Sizer Policy Production
Threshold %
       CPU 85%
      RAM 85%
      SSD 85%
     HDD 85%
Minimum Networking speed 10G
Minimum SSD per node 1
Minimum socket per node 1
CVM : vCPU:pcore 2
Compression % (default) 30%
ECX OFF

Cluster settings UI in Sizer:

What is an Expert Library ?

Expert library is exactly same as the cluster profiles except that instead of the users defining the cluster profiles themselves, these profiles are created by experts in their respective areas.

The Expert library of cluster profiles is  listed by workloads.  For ex: Database experts know best what an Oracle cluster should look like and apply the settings accordingly. Similarly, there are libraries for VDI, Files , General purpose Server Virtualisation etc.

SEs and partners would just need to apply these libraries during creating a sizing for the respective workload/s  from the cluster settings page (explained in detail in next section).

Expert Library is a separate tab next to the user defined cluster profile in Sizer.

How to use Cluster template/Expert libraries ? 

While creating a sizing, on the cluster settings page, click on ‘Apply cluster template’ which pops up the list of user defined profiles and expert libraries. SEs can select any and have a quick view of the set attributes and hit apply. This will automatically select values in the cluster settings screen for the scenario which is being sized.

Similarly, as shown in the below image, using the Save template button, a new Profile can be saved from the cluster settings page of a scenario being sized. This becomes handy for leveraging the work done in manually setting the filters for current sizing for future reference for a new sizing.

For example: If an SE is sizing for a customer BestBuy which prefers certain NX models , SE would need to select only those models in platform list in the Cluster settings page for the given scenario. At the same time, they can save this as a template (say naming it as BestBuy template) and can apply every time when sizing for BestBuy opportunity.

List/View/Apply Cluster template while creating a sizing:

 

 

A demo on creating and using templates in Sizer:

 

Posted on

Feb 2023 (HW Template)

What is a HW Template in Sizer ? 

HW (Hardware) templates are pre-defined node configurations that can be created and saved by user. These will appear in their list of saved HW templates and can be used/applied while current or future sizings.

Similar to Cluster templates, HW template is a user-defined settings, only difference being HW templates allows to create exact specific HW config whereas Cluster templates provides attributes or settings guiding the sizer to pick up best configuration.

How can one create a HW Template?

There are two ways to define your fixed configuration of node via HW templates:

One – Upon landing to sizer (my scenario) dashboard, users can click on “Templates” and select HW Templates which opens up the HW configuration page ( similar to Manual mode page) where they can choose the vendor / platform and everything else (CPU/MeM/SSD/HDD/NICs/GPU)

Second – In order to save the effort, users can also create the HW templates while in the midst of a current sizing. While in a sizing exercise, in manual mode, when users would be defining a configuration , they can also quickly save it as a HW template in one additional step by giving it a name (either for a workload or by customer for example: HW Template-Databases, or HW Template-BestBuy)

How to use the HW templates ?

The objective for HW template is that you create once and use every time for that particular situation ( by workload or for a particular customer or any other scenario).

After creating the workload, when users are presented the Cluster settings/Sizing options screen, it shows them the option of HW Templates which can be applied for that sizing. This is in along with the other Cluster / SW settings they would like to use for that scenario. Sizer , then, iterates through the different solutions meeting the workload criteria and the cluster settings but restricting itself to only the HW Template(s) that the user selected for that sizing. Sizer would then recommend the right number of nodes (of the selected HW template(s)) meeting all the workload and cluster settings criteria.

How is HW Templates useful ? 

The biggest advantage for HW Templates is that it optimizes the sizing activity for situations where their is a fixed BOM involved. That could be whether based on a customers existing server farm or a Service Provider or a reseller/distributor with a fixed BOM(like Arrow with custom built HPE configs) or simply SE preferred config for a given workload.

Another use case can be for upgrades. SEs can define the NG8 equivalent of their older HW upgrades(like NX G5) and always map the existing cluster on older models to a mapped one via Templates.

Once the config is fixed, Sizer then comes up with number of nodes avoiding any back and forth for number of nodes of that BOM to meet the criteria.

Also, for future sizings for the similar criteria, users need not do all the manual exercise all over again (and even run the risk of manual errors) but simply import the HW Template for that criteria and get uniform/consistent sizing in one single click.

While creating the fixed BOM configuration, HW Templates also ensures all the platform rules are adhered to.   Additionally, after importing the templates, users can edit/modify configs and tweak it from the original template and before going ahead with the sizing, can save the edited template with a different name.

 

Here is a short demo on HW Templates, how to create and use : 

Posted on

January 2023

Major highlights for January : 
Upgrade sales flow :
  • Now upgrade flow also supports HPE DX
  • You can recreate existing cluster with older DX configs (just as NX)
  • This will be a manual exercise though (like NX) but auto recreate will be coming soon
Imported workload enhancements:
Storage for powered Off VMs
  • Option to consider the storage for powered Off VMs (even if sized for only powered ON VMs)
  • This is applicable across Collector/RVTool/Insights imports
  • Checked by default but can opt-out
Retain cluster mapping
  • This will retain the source cluster and the VMs mapped to accordingly to the sizing
  • Instead of consolidating all VMs to default Cluster-1
  • Not checked by default so need to opt-in

Posted on

November 2022

Here are the latest release with the following updates on SW/Platform:
  • Snapshots: RPO frequency updates for AOS 6.5.1
  • One-hour supported for up to 136TB for hybrid nodes  plus other rule changes
  • Adding 2% buffer for max capacity restrictions
  • 15.36TB NVMe SSDs support / Storage Capacity calculator
  • Lenovo Rule: Limit CPUs with wattage <=205W
Licensing update:
  • NCI : Handling the free TiBs of Files / Objects with the NCI license (also for Legacy)
  • EUC : Handling free GBs per user for Files with per user VDI license
Usability :
  • My scenarios dashboard:  Dynamically adjustable columns
  • Makes search by name easier
  • Including shared scenario dashboard
  • Partner login:
  • Updating deal registration-based validation for partner flow
  • Allow sizing for all approved deal registration ids
  • NCP bundle:
  • Highlighting individual licenses across tiers
Posted on

October 2022

Here are the updated on two new features:
TCO in Proposals:
  • A new section on TCO in the Proposal deck
  • It has charts comparing the total cost of the solution on NX vs AWS
  • It is based on pre-build scenarios of standard sizes that is closest to the one being sized
  • For example: for a sizing of 450 VDI users, it will pick a pre-build scenario of 500 VDI users
  • It’s an optional and editable section in the proposal ,and hope it serves as a good conversation starter
  • The section also has guide on where and how to reach out to TCO team for a more detailed analysis/questions
CVM modes :
  • To address wide range of requests and different use cases, introduced two modes for allocating CVM cores
  • Regular mode: It is the current behaviour (and default option) so no changes
  • Performance mode: If selected, it doubles up the cvm cores
  • For example: for a regular VDI sizing, sizer allocates 4 cores , but if performance mode, will allocate 8 cores
  • Particularly helpful in addressing the resources required for certain configs, for example: NVMe nodes
  • NVMe nodes are used even for non-performance critical use case due to price parity with SSDs so regular allocation should be fine
  • In cases where it is used for performance , selecting the mode will ensure higher cores are provisioned
Other updates:
  • Updated HPE DX HW pricing for budgetary quotes (for DX BOM)
  • Lenovo product updates
Posted on

August 2022

Here are the major highlights listed below :
Products:
  • Proposals – includes NG8 and HPE DXG10Plus models
  • HPE DX HW BOM – pricing update for budgetary quote
  • NCM Cloud add-on in Solutions
  • NUS add-on to NCI update (can add to all tiers)
  • NC2(AWS) – heterogeneous cluster support (mixing of different server types)
  •  NVME nodes CVM update (parity with SSD node type) – regular use case
Platform:
  • NX NG8 NIC changes : support for configurable LOM(AIOM)
  • Fujitsu – additional memory configs support