Nutanix AOS 5.5 delivers 1M read IOPS from a single VM, but what about 70/30 read/write?

I recently wrote Nutanix AOS 5.5 delivers 1M IOPS from a single VM, but what happens when you vMotion which showed the impact of a vMotion was around -10% for a period of approx. 3 seconds before read performance resumed back to pre-migration levels.

In this post I will be addressing the question about performance for a single VM with a more realistic 70% read, 30% write IO profile which was performed using an 8k IO size and what the impact is during and after a live migration.

While not surprising to Nutanix customers, this result shows a maximum starting baseline of 436K random read and 187k random write IOPS and immediately following the migration performance reduced to 359k read and 164k write IOPS before achieving greater performance than the original baseline @ 446k read and 192k IOPS within a few seconds.

So in comparison to 100% random read which achieved just over 1 million 8k IOPS, the 70/30 mix achieves in the ballpark of 600k IOPS which is very respectable. Not bad for a platform which Nutanix competitors continue to describe as only being good for VDI. Considering even the largest array from a leading all flash SAN vendor is only advertising performance in the hundreds of thousand random read range, it shows Nutanix unique hyper-converged architecture can achieve higher performance than a monolithic all flash array from a single VM.

This shows that with the unique Nutanix Acropolis Distributed Storage Fabric, very high performance at low latency can be achieved with real world IO patterns even during and after live migrating the virtual machine across a distributed platform.

This result is further evidence of the efficiency of Nutanix Acropolis Hypervisor, AHV (which is included at no additional charge with AOS) as well as the IO path running in user space (not the much hyped in-kernel). This is in part thanks to AHV Turbo Mode which optimised the IO path which was announced at .NEXT 2017 in Washington. In addition to these excellent levels of performance, they can be sustained even when using data protection features such as snapshots as shown in recent post I wrote about Nutanix X-ray tool where I used the Snapshot impact scenario to compare Nutanix AHV and a leading hypervisor and SDS product. If you don’t have time to read the post, in short, the Nutanix competitors performance degraded as snapshots were taken while Nutanix AHV’s performance remained consistent which is essential for real world scenarios, especially with business critical applications.

With Nutanix unique ability to scale out performance using storage only nodes, even higher performance can be achieved without modification to the virtual machine to applications which gives Nutanix further advantage over the competition.

Nutanix data locality ensures optimal performance by ensuring new data is always local to the VM and cold data can remain remote indefinitely while only hot data will be migrated locally if/when required at a 1MB granularity. This translates to intelligent data locality and not brute force locality as it is frequently mistaken to be.

Back to Part 1

Nutanix AOS 5.5 delivers 1M IOPS from a single VM, but what happens when you vMotion?

For many years Nutanix has been delivering excellent performance across multiple hypervisors as well as hardware platforms including the native NX series, OEMs (Dell XC & Lenovo HX) and more recently software only options with Cisco and HPE.

Recently I tweeted (below) showing how a single virtual machine can achieve 1 million 8k random read IOPS and >8GBps throughput on AHV, the next generation hypervisor.

While most of the response to this was positive, the usual negativity came from some competitors who tried to spread fear, uncertainty and doubt (FUD) about the performance including claims it was not sustainable during/after a live migration (vMotion) and that is does not demonstrate the performance of the IO path.

Let’s quickly cover of the IO path discussion of in-kernel vs a controller VM.

To test the IO path, in the case of Nutanix, via the Controller VM, you want to eliminate as many variables and bottlenecks as possible. This means a read/write test is not valid as writes are dependant on factors such as the network. As this was one a node using NVMe, the bottleneck would quickly become the network and not the path between the user VM and controller VM.

I’ve previously tweeted (below) showing an example of the throughput capabilities of SATA SSD, NVMe and 3DxPoint which clearly shows the network is the bottleneck with next generation flash.

I’ve also responded to 3rd party FUD about Nutanix Data locality with a post which goes in depth about Nutanix original & unique implementation of Data Locality which is how Nutanix minimises its dependancy on the network to deliver excellent performance.

So we are left with read IO to actually test and possibly stress the IO path between a User VM and software defined storage, be that in-kernel or in user space which is where the Nutanix CVM runs.

The tweet showing >1 million 8k random read IOPS and >8GBps throughput shows that the IO path of Nutanix is efficient enough to achieve this at just 110 micro (not milli) seconds.

The next question from those who try to discredit Nutanix and HCI in general is what happens after a vMotion?

Let me start by saying this is a valid question, but even if performance dropped during/after a vMotion is it even a major issue?

For business critical applications, it is common for vendors to recommend DRS should/must rules to prevent vMotion exception for in the event of maintenance or failure regardless of the infrastructure being traditional/legacy NAS/SAN or HCI.

With a NAS/SAN, the best case scenario is 100% remote IO where as with Nutanix this is the worse cast scenario. Let’s assume business as usual on Nutanix is 1M IOPS and during a vMotion and for a few mins after that performance dropped by 20%.

That would still be 800k IOPS which is higher than what most NAS/SAN solutions can delivery anyway.

But the fact is, Nutanix can sustain excellent performance during and after a vMotion as demonstrated by the video below which was recorded in real time. Hint: Watch the values in the putty session as these show the performance as measured at the guest level which is what ultimately matters.

Credit for the video goes to my friend and colleague Michael “Webscale” Webster (VCDX#66 & NPX#007).

The IO dropped below 1 million IOPS for approx 3 seconds during the vMotion with the lowest value recorded at 956k IOPS. I’d say an approx 10% drop for 3 seconds is pretty reasonable as the performance drop is caused by the migration stunning the VM and not by the underlying storage.

Over to our “friends” at the legacy storage vendors to repeat the same test on their biggest/baddest arrays.

Not impressed? Let’s see what 70/30 read/write workload performs!

A TCO Analysis of Pure FlashStack & Nutanix Enterprise Cloud

In helping to prepare this TCO with Steve Kaplan here at Nutanix, I’ll be honest and say I was a little surprised at the results.

The Nutanix Enterprise Cloud platform is the leading solution in the HCI space and it while it is aimed to deliver great business outcomes and minimise CAPEX,OPEX and TCO, the platform is not designed to be “cheap”.

Nutanix is more like the top of the range model from a car manufacturer with different customer requirements. Nutanix has options ranging from high end business critical application deployments to lower end products for ROBO, such as Nutanix Xpress model.

Steve and I agreed that our TCO report needed to give the benefit of the doubt to Pure Storage as we do not claim to be experts in their specific storage technology. We also decided that as experts in Nutanix Enterprise Cloud platform and employees of Nutanix, that we should minimize the potential for our biases towards Nutanix to come into play.

The way we tried to achieve the most unbiased view possible is to give no benefit of the doubt to the Nutanix Enterprise Cloud solution. While we both know the value that many of the Nutanix capabilities have (such as data reduction), we excluded these benefits and used configurations which could be argued at excessive/unnecessary such as vSphere or RF3 for data protection:

  1. No data reduction is assumed (Compression or Deduplication)
  2. No advantage for data locality in terms of reduced networking requirements or increased performance
  3. Only 20K IOPS @ 32K IO Size per All Flash Node
  4. Resiliency Factor 3 (RF3) for dual parity data protection which is the least capacity efficient configuration and therefore more hardware requirements.
  5. No Erasure Coding (EC-X) meaning higher overheads for data protection.
  6. The CVM is measured as an overhead with no performance advantage assumed (e.g.: Lower latency, Higher CPU efficiency from low latency, Data Locality etc)
  7. Using vSphere which means Nutanix cannot take advantage of AHV Turbo Mode for higher performance & lower overheads

On the other hand, the benefit of the doubt has been given to Pure Storage at every opportunity in this comparison including the following:

  1. 4:1 data reduction efficiency as claimed
  2. Only 2 x 10GB NICs required for VM and Storage traffic
  3. No dedicated FC switches or cables (same as Nutanix)
  4. 100% of claimed performance (IOPS capability) for M20,M50 and M70 models
  5. Zero cost for the project/change control/hands on work to swap Controllers as the solution scales
  6. IOPS based on the Pure Storage claimed average I/O size of 32K for all IO calculations

We invited DeepStorage and Vaughn Stewart of Pure Storage to discuss the TCO and help validate our assumptions, pricing, sizing and other details. Both parties declined.

Feedback/corrections regarding the Pure Storage sponsored Technical Report by DeepStorage was sent via Email, DeepStorage declined to discuss the issues and the report remains online with many factual errors and an array (pun intended) of misleading statements which I covered in detail in my Response to: DeepStorage.net Exploring the true cost of Converged vs Hyperconverged Infrastructure

It’s important to note that the Nutanix TCO report is based on the node configuration chosen by DeepStorage with only one difference: Nutanix sized for the same usable capacity, but went with an All Flash solution because comparing hybrid and all flash is apples and oranges and a pointless comparison.

With that said, the configuration DeepStorage chose does not reflect an optimally designed Nutanix solution. An optimally designed solution would likely result in fewer nodes by using 14c or 18c processors to match the high RAM configuration (512GB) and different (lower) capacity SSDs (such as 1.2TB or 1.6TB) which would deliver the same performance and still meet the capacity requirements which would result in a further advantage in both CAPEX, OPEX and TCO (Total Cost of Ownership).

The TCO shows that the CAPEX is typically in the favour of the Nutanix all flash solution. We have chosen to show the costs at different stages in scaling from 4 to 32 nodes – the same as the DeepStorage report. The FlashStack product had slightly lower CAPEX on a few occasions which is not surprising and also not something we tried to hide to make Nutanix always look cheaper.

One thing which was somewhat surprising is that even with the top of the range Pure M70 controllers and a relatively low IO per VM assumption of 250, above 24 nodes the Pure system could not support the required IOPS and an additional M20 needed to be added to the solution. What was not surprising is in the event an additional pair of controllers and SSD is added to the FlashStack solution, that the Nutanix solution had vastly lower CAPEX/OPEX and of course TCO. However, I wanted to show what the figures looked like if we assume IOPS was not a constraint for Pure FlashStack as could be the case in some customer environments as customer requirements vary.

PureVNutanixComparisonWithLowerIOPS

What we see above is the difference in CAPEX is still just 14.0863% at 28 nodes and 13.1272% difference at 32 nodes in favor of Pure FlashStack.

The TCO, however, is still in favor of Nutanix at 28 nodes by 8.88229% and 9.70447% difference at 32 nodes.

If we talk about the system performance capabilities, the Nutanix platform is never constrained by IOPS due to the scale out architecture.

Based on Pure Storage advertised performance and a conservative 20K IOPS (@ 32K) per Nutanix node, we see (below) that Nutanix IO capability is always ahead of Pure FlashStack, with the exception of a 4 node solution based on our conservative IO assumptions. In the real world, even if Nutanix was only capable of 20K IOPS per node, the platform vastly exceeds the requirements in this example (and in my experience, in real world solutions) even at 4 node scale.

PurevsNTNXIOPS

I’ve learned a lot, as well as re-validated some things I’ve previously discovered, from the exercise of contributing to this Total Cost of Ownership (TCO) analysis.

Some of the key conclusions are:

  1. In many real world scenarios, data reduction is not required to achieve a lower TCO than a competing product which leverages data reduction.
  2. Even the latest/greatest dual controller SANs still suffer the same problems of legacy storage when it comes to scaling to support capacity/IO requirements
  3. The ability to scale without rip/replace storage controllers greatly simplifies customers sizing
  4. Nutanix has a strong advantage in Power, Cooling, Rack Space and therefore helps avoid additional datacenter related costs.
  5. Even the top of the range All Flash array from arguably the top vendor in the market (Pure Storage) cannot match the performance (IOPS or throughput) of Nutanix.

The final point I would like to make is the biggest factor which dictates the cost of any platform, be it the CAPEX, OPEX or TCO is the requirements, constraints, risks and assumptions. Without these, and a detailed TCO any discussion of cost has no basis and should be disregarded.

In our TCO, we have detailed the requirements, which are in line with the DeepStorage report but go further to make a solution have context. The Nutanix TCO report covers the high level requirements and assumptions in the Use Case Descriptions.

Without further ado, here is the link to the Total Cost of Ownership comparison between Pure FlashStack and Nutanix Enterprise Cloud platform along with the analysis by Steve Kaplan.