Category Archives: Acropolis

MS Exchange on Nutanix Acropolis Hypervisor (AHV)

Posted on by
Reply

While Virtualization of MS Exchange is now common across multiple hypervisors it continues to be a hotly debated topic. The most common objections being cost (CAPEX), the next being complexity (which translates to CAPEX & OPEX) and the third being that virtualization adds minimal value as MS Exchange provides application level high availability. The other objection I hear is Virtualization isn’t supported, which always makes me laugh.

In my experience, the above objections are typically given in the context of a dedicated MS Exchange environment, which in that specific context some of the points have some truth, but the question becomes, how many customers run only MS Exchange? In my experience, None.

Customers I see typically run tens, hundreds even thousands of workloads in their datacenters so architecting silos for each application is what actually leads to cost & complexity when we think outside the box.

Since most customers have virtualization and want to remove silos in favour of a standarized platform, MS Exchange is just another Business Critical Application which needs to be considered.

Let’s discuss each of the common objections and how I believe Acropolis + Nutanix XCP addresses these challenges:

Microsoft Support for Virtualization

For some reason, there is a huge amount of FUD regarding Microsoft support for Virtualization (other than Hyper-V), but Nutanix + Acropolis is certified under the Microsoft Server Virtualization Validation Program (SVVP) and runs on block storage via iSCSI protocol, so Nutanix + Acropolis is 100% supported for MS Exchange as well as other workloads like Sharepoint & SQL.

Cost (CAPEX)

Unlike other hypervisors and management solutions, Acropolis and Acropolis Hypervisor (AHV) come free with every Nutanix node which eliminates the licensing cost for the virtualization layer.

Acropolis management components also do not require purchase or installation of Tier 1 database platforms, all required management components are built into the distributed platform and scaled automatically as clusters are expanded. As a result, even licenses for Windows operating system are not required.

As a result, Nutanix + Acropolis gives Exchange deployments all the Virtualization features (below) which provide benefits at no cost.

  • High Availability & Live Migration
  • Hardware abstraction
  • Performance monitoring
  • Centralized management

Complexity (CAPEX & OPEX)

Nutanix XCP + Acropolis can be deployed in a fully optimal configuration from out of the box to operational in less than 60 minutes. This includes all required management components which are automatically deployed as part of the Nutanix Controller VM (CVM). For single cluster environments, no design/installation is required for any management components, and for multiple-cluster environments, only a single virtual appliance (PRISM Central) is required for single pane of glass management across all clusters.

Acropolis gives Exchange deployments all the advantages of Virtualization without:

  • Complexity of deploying/maintaining of database server/s to support management components
  • Deployment of dedicated management clusters to house management workloads
  • Having onsite Subject Matter Experts (SMEs) in Virtualization platform/s

Virtualization adds minimal value

While applications such as Exchange have application level high availability, Virtualization can further improve resiliency and flexibility for the application while making better use of infrastructure investments.

The Nutanix XCP including Acropolis + Acropolis Hypervisor (AHV) ensures infrastructure is completely abstracted from the Operating System and Application allowing it to deliver a more highly available and resilient platform.

Microsoft advice is to limit the maximum compute resources per Exchange server to 24 CPU cores and 96GB RAM. However with CPU core counts continuing to increase, this may result in larger numbers of servers being purchased and maintained where an application specific silo is deployed. This would lead to increased datacenter and licensing costs not to mention operational overhead of managing more infrastructure. As a result, being able to run Exchange alongside other workloads in a mixed environment (where contention can easily be avoided) reduces the total cost of infrastructure while providing higher levels of availability to all workloads.

Virtualization allows Exchange servers to be sized for the current workload and resized quickly and easily if/when required which ensures oversizing is avoided.

Some of the benefits include:

  • Minimizing infrastructure in the datacenter
  • Increasing utilization and therefore value for money of infrastructure
  • Removal of application specific silos
  • Ability to upgrade/replace/performance maintenance on hardware with zero impact to application/s
  • Faster deployment of new Exchange servers
  • Increase availability and provide higher fault tolerance
  • Self-healing capabilities at the infrastructure layer to compliment application level high availability
  • Ability to increase Compute/Storage resources beyond that of the current underlying physical server (Nutanix node) e.g.: Add storage capacity/performance

The Nutanix XCP Advantages (for Exchange)

  • More usable capacity

With features such as In-Line compression giving between 1.3:1 and 1.7:1 capacity savings & Erasure Coding providing up to a further 60% usable capacity, Nutanix XCP can provide more usable capacity than RAW while providing protection from SSD/HDD and entire server failures.

In-Line compression also improved performance of the SATA drives, so its a Win/Win. Erasure coding (EC-X) stores data in a more efficient manner which allows more data to be served from the SSD tier, also a Win/Win.

  • More Messages/Day and/or Users per physical CPU core

With all Write I/O serviced by SSD the CPU WAIT time is significantly reduced which frees up the physical CPU to perform other activities rather than waiting for a slow SATA drive to respond. As MS Exchange is CPU intensive (especially from 2013 onwards) this means more Messages per Day and/or Users can be supported per MSR VM compared to physical servers.

  • Better user experience

As Nutanix XCP is a hybrid platform (SSD+SATA), newer/hotter data is serviced by the SSD tier which means faster response times for users AND less CPU WAIT which also helps further increase CPU efficiencies, again leading to more Messages/Day and/or Users per CPU core.

Summary:

With Cost (CAPEX), Complexity (CAPEX & OPEX) and supportability issues well and truly addressed and numerous clear value adds, running a business critical application like MS Exchange on Nutanix + Acropolis Hypervisor (AHV) will make a lot of sense for many customers.

Acropolis Hypervisor (AHV) & non-uniform node CPU generations

Posted on by
2

For those of you familiar with VMware vSphere’s Enhanced vMotion Compatibility (EVC) feature, you might be wondering how non-uniform CPU generations are handled in an Acropolis Hypervisor (AHV) environment.

Well, as with most things Nutanix, the answer is simple.

NOS 4.5 automatically detects and configures the lowest common CPU generation as the baseline on a per cluster basis.

The following diagram shows how it works:

AHVEVC2

As we can see, we have a four node Acropolis cluster with 3 different CPU generations. Acropolis detects Sandy Bridge as the lowest common denominator and ensures VMs on all nodes are only exposed the Sandy Bridge CPU capabilities.

This ensures Live migration capabilities are maintained across the cluster.

Note: As with vSphere’s EVC, VMs still benefit from higher clock rates and performance from newer generation CPUs, they just don’t have all CPU capabilities exposed, so don’t be fooled into thinking your newer/faster CPUs are wasted in a mixed environment.

NOS 4.5 Delivers Increased Read Performance from SATA

Posted on by
2

In a recent post I discussed how NOS 4.5 increases the effective SSD tier capacity by performing up-migrations on only the local extent as opposed to both RF copies within the Nutanix cluster. In addition to this significant improvement in usable SSD tier, in NOS 4.5 the read performance from the SATA tier has also received lots of attention from Nutanix engineers.

What the Solutions and Performance Engineering team have discovered and been testing is how we can improve SATA performance. Now ideally the active working set for VMs will fit within the SSD tier, and the changes discussed in my previous post dramatically improve the chances of that active working set fitting within the SSD tier.

But there are situation when reads to cold data still need to be serviced by the slow SATA drives. Nutanix uses Data Locality to ensure the hot data remains close to the application to deliver the lowest latency and overheads which improve performance, but in the case of SATA drives and the fact data is infrequently accessed from SATA means that reading from remote SATA drives can improve performance especially where the number of local SATA drives is limited (in some cases to only 2 or 4 drives).

Most Nutanix nodes have 2 x SSD and 4 x SATA so best case you will only see a few hundred IOPS from SATA as that is all they are physically capable of. To get around this issue.

NOS 4.5 introduces some changes to the way in which we select a replica to read an egroup from the HDD tier. Periodically NOS (re)calculate the average IO latencies of the all the replicas of a vdisk’s (replicas which have the vdisk’s egroups). We use this information to choose a replica as follows:

  1. If the latency of the local replica is less than a configurable threshold, read from the local replica.
  2. If the latency of the local replica is more than a configurable threshold, and the latency of the remote replica is more than that of the local replica, prefer the local replica.
  3. If the latency of the local replica is more than a configurable threshold and the remote replica is lower than the configurable threshold OR lower than the local copy, prefer the remote replica.

The diagram below shows an example of where the VM on Node A is performing random reads to data A and shortly thereafter data C. When requesting reads from data A the latency is below the threshold but when it requests data C, NOS detects that the latency of the local copy is higher than the remote copy and selects the remote replica to read from. As the below diagram shows, one possible outcome when reading multiple pieces of data is one read is served locally and the other is serviced remotely.

remotesatareads2

Now the obvious next question is “What about Data Locality”.

Data Locality is being maintained for the hot data which resides in SSD tier because reads from SSD are faster and have lower overheads on CPU/Network etc when read locally due to the speed of SSDs. For SATA reads which are typical >5ms the SATA drive itself is the bottleneck not the network, so by distributing the Reads across more SATA drives even if they are not local, results in better overall performance and lower latency.

Now if the SSD tier has not reached 75% all data will be within the SSD tier and will be served locally, the above feature is for situations where the SSD tier is 75% full and data is being tiered to SATA tier AND random reads are occurring to cold data OR data which will not fit in the SSD tier such as very large databases.

In addition NOS 4.5 detects if the read I/O is random or sequential, and if its sequential (which SATA performance much better at) then the up-migration of data has a higher threshold to meet before being migrated to SSD.

The result of these algorithm improvements (and the increased SSD tier effective capacity discussed earlier) and Nutanix In-line compression is higher performance over larger working sets which also exceed the capacity of the SSD tier.

Effectively NOS 4.5 is delivering a truly scale out solution for read I/O from SATA tier which means one VM can be reading from potentially all nodes in the cluster ensuring SATA performance for things like Business Critical Applications is both high and consistent. Combine that with NX-6035C storage only nodes, this means SATA read I/O can be scaled out as shown in the below diagram without scaling compute.

ScaleOutRemoteReads

 

As we can see above, the Storage only Nodes (NX-6035C) are delivering additional performance for read I/O from the SATA tier (as well as from the SSD tier).