Dell EMC Powerstore appliance Review Introduction
In this blog I will discuss a few of the new PowerStore’s features and selling points. If we go through all of this, we would see that some other products might fall away or just be replaced by PowerStore. These include products like VxRAIL, VxFlex, etc… You can also see this as a kind of Hyper-Converged solution, but it is called a midrange storage solution. We must also remember all the other midrange storage products like EqualLogic, Compellent XtremeIO probably VNXe will now be replaced by PowerStore. So we finally have an answer to all the questions of what will happen to these legacy products…
Let us have a look at some features that make PowerStore stand out from other vendors or even some already owned Dell EMC equipment. Respect for Daniel Cummins (Chief Architect – PowerStore) and he’s team members for a great product as they addressed a bunch of requirements in the current cloud space with this storage array. With this product, I see that it can compete against the likes NetApp, Pure Storage, HPe arrays, etc, and what is more, you can also use it as a Converged and or Hyper-Converged system.
Different models – two deployment modes
You will get your “Standard deployment or PowerStore T” and “Hypervisor deployment or PowerStore X”. I mentioned earlier that we can use these to replace either your current VxRail or VxFlex products. Now in here, you see that by adopting a standard deployment it kind of does the same as VxFlex or new “PowerFlex” with old school block, etc.. storage assigned to anything. And by adopting a Hypervisor deployment you can replace or just add to most of your current VxRail clusters. Dell EMC claims that VxRail is built for general purpose and PowerStore is built for specialized apps. Just thinking about the logic here my guess would be that the X model is VMware vSAN in other words “VxRail” and the T model build on the ScaleIO OS (VxFlex) they bought a few years ago with a bunch of names and add-ons to make it stand out as something new. But by doing a bit of research and by saying this I am wrong in the sense that VxRAIL uses software-defined storage vSAN and VxFlex is using software-defined storage called VxFlex OS (formerly ScaleIO). With the new PowerstoreOS, it is a bit different as there are no underlying SDS solution and now they can call it “Native and or Containers”
When I play with the demo, I see that the Hypervisor model possibly uses plain old block storage inside the same appliance called “Containers” and presented as vVOLS to VMware and vSAN does not play a part here. The same goes for the standard model where you assign block to whatever environment…
So differentiating between Powerstore, VxFlex, and VxRail. You need not use an SDS solution to see the disks inside, let’s say a server anymore like with VxFLEX and VxRAIL. Now it is swapped around where you have a storage solution with a hardware server that is integrated if you understand my logic.
Features and selling points
End to end NVMe
This NVMe does not support NVMe-oF (over Fabrics) yet, I am sure it will later on. Dell EMC uses a 2U form factor with dual controllers. Each array has 25 2.5” bays (up to 23 usable for storage depending on model), supporting drive capacities ranging from 375GB to 750GB in Optane SSDs and 1.92TB to 15.36TB NVMe SSDs. All good and well as it is all in one appliance. But if you think about it, then few other vendors can do this as well. The VxRail P580N can also do this but we must remember the PowerStore is an all in one midrange storage solution with features from some of their other products.
Scale up and scale out
Ok so Dell EMC claims that they can scale up and scale-out. On the “Hypervisor” models you can scale out and on the “Standard” models you scale-out which is good as with all Hyper-converged solutions (as we can classify this as) these days we would at least expect this. The scale-up looks like it can be done by adding up to three enclosures of storage and resources with the standard model only. This makes sense as you would use the hypervisor model kind of like a VxRail. If I am not confused, then it looks like the standard model scale-up is done with SAS. Correct me if I am wrong but this is not end to end NVMe anymore?
Innovative AppsON capability lets you run
virtualized workloads directly on the array
I am going to try and break down what AppsON is. At this stage, to me, it just sounds like one of those names that generate sales, but no one knows exactly what it is doing.
So I will start by saying it is is mostly VMware dependent with on-appliance native storage and simplified. You can also look at this as VMware SDDC running on VMware cloud foundation – a software-defined data center where you can move your workloads between devices/clouds.
So in essence you get a view in PowerStore/VMware with more integration, management, and apps running on with vVOLS…. wait for it… “The PowerStore Operating System” made of a group of… “Containers – and then called Micro-services”. Let me explain why I emphasized this. So AppsON automatically runs on containers. If we drill down to what is behind the scenes, we see that you have your let’s say traditional midrange Unity storage behind, where you would typically deploy Pools, LUNs and a NAS with file systems. Now you have a similar scenario but instead of calling a LUN a LUN, it is a “volume” now. And instead of calling a Pool a Pool, it is a “Container” now and a group of containers is micro-services now… So if you have an out of the box PowerStore X you automatically have a few containers where your apps are on, these pools aka containers are then called “micro-services” – it makes me think of the old Austin Powers movies with Dr evil and mini-me :). Don’t get me wrong, this is still an awesome product, but let’s call a spade a spade. You will be able to build the same solution with a VxRail and a bunch of other vendors but some names will differ and because containers are in now let us just call it that. Another thing is that your Apps are on the same appliance communicating natively with the storage and there are now underlying SDS needed.
As we discussed above regarding the LUNs and Pools or Volumes and Containers, I will look a bit at the admin side without diving into the technical hardware. As with all other midrange storage arrays like Unity etc, you create your volumes and group it into volume groups if needed – the same principle than all other Dell EMC midrange storage. You can then during creation let these LUNs\Volumes choose automatically on which appliance it will run or manually set it as below. You can also Specify a protection policy with integrated snapshots and set the performance required.
When you create a container, I must add that there is a bit more integration than a normal pool on Unity. Although the exact same can be manually done on a Unity array. Containers/Pools are created as I will show you below and the one difference from a normal pool is that you need to add it in VMware as a vVol.
With the NAS server and file Systems you can create, it is pretty much exactly like on the Unity. You create your NAS with a specified IP and file system afterward. I would have liked to see them switch the two options around as a NAS is a prerequisite for creating a file system.
Dedupe and compression
So we get build in, inline, always on and 4:1 guaranteed deduplication and compression on this appliance. This is always useful and they say that there is no performance price to pay. So from my experience with Unity all-flash arrays, I must add that this is true.
Dell EMC runs a future proof program which includes “Anytime upgrade” In this program, you can choose consumption models for your PowerStore solution:
Next-Gen: Upgrade appliance nodes (controllers) to the next generation
• Higher Model: Upgrade to more powerful nodes within the current
• Scale-Out: Apply a discount to expand your environment with a second
system equal to the current model.
We can add these models changed non-disruptively. Dell Technologies On-Demand provides a range of options to easily purchase and scale storage — as you grow it, as you use it, or as a service. In environments where capacity demands are cyclical or variable, usage-based consumption models deliver distinct advantages of something that is needed on Dell EMC’s storage portfolio.
This is a dominant move from Dell EMC and now we know what will replace all the other midrange overlapping storage arrays they have. We can also see that this can even replace some current HCI solutions and will surely be used as a storage and resource component in the next-gen VxBlock or rather PowerOne models. To be honest with you, I always respected Dell and EMC and by the looks of it they came up with a powerful solution that ultimately has the potential to be a market leader again.
Check out the whitepaper for more info
“Some images or content may be the property of Dell Technologies“