VMware vSAN alternatives and competitors

It's really about simplifying my infrastructure and getting myself to be cloud-enabled in a hybrid infrastructure so that I can move in and out of the cloud. That way, I can have a very easy-to-manage, intelligent infrastructure so I no longer have to worry about level one and level two calls. The storage is intelligent, it's taken care of itself. For example, if I want full-stack management, or I want to know if my virtual machines are being bad actors or noisy neighbors, or the firmware on my servers or hosts, or my hardware is all in good shape or if it has a predictive failure — I want to know about it and I want it to open up a ticket for me so that the part can be delivered to me and I don't have to open up a ticket. 

Here's a good scenario: Years ago I used to manage large data centers. I started out with 800 servers, I had 16 people. When I left there, I had 2,900 servers and eight people. That is the reality of IT staff. They're constantly being compressed and asked to do more with less. In order to do that you have to have an intelligent infrastructure. You have to have a methodology that allows you to be able to supply your user community with the services that they need consistently with a smaller staff. That is what really drives DACI and HCI.

Our clients come in all sizes, from small, medium, and large-sized businesses.

Nimble gives time back to our clients. They know it's secure. They know that it's self intelligence software. They know that it's going to report to them and let them know if there's an issue — it's going to give them predictive data. It's going to give them a full stack of data. What a lot of them couldn't see before was what's going on with the virtual environment. Now they can see that because it tells them. So not only is it storage, it's telling them, "Hey, here's how your virtual machines are acting, and oh, by the way, you need some more memory on this server." You don't expect that from any other storage. I've only seen it with Nimble. It gets them out of that firefighting mode. Now they can actually go do the things that their customers are asking them to do. That's how it changed their life from day to day for the business.

What's really intriguing about it is they truly took everything that was required to run on the host, typically in an HCI solution, and ran it out of the storage. 

You don't have to have any virtual controllers or any of that management that takes away from the process or a memory that you would use for your virtualization. You don't have to take any of your resources away — it's brilliant.

The nice thing about it is that it truly integrates with the cloud, so I can move my volumes bi-directionally in and out of the cloud. It has intelligent replication to the point where I can replicate to two different sites and to the cloud. 

There's a three, two, one, zero rule. That means I need three copies of my data on two different types of media and I need one of them to be protected from ransomware. I need it to be immutable. The zero stands for zero errors; I need to know that my backup was successful and I've actually tested it and can say that I can restore from it. A lot of people miss that point. 

The intelligence of the solution, in general, is great. There are other great features like Triple+ parity RAID. They are on par, or better than any of the other storage vendors out there for a much more reasonable price, but what really puts them over the top for customers is the intelligence.

Every time a vendor buys someone else, and HP is no different than anyone else, they usually mark it up. When they bought Nimble, I said, "Oh man, that's it. Nimble's done. They're going to screw it up." And they didn't. They actually invested a lot of time and a lot of resources to make it better and to take what was good inside of Nimble, including the InfoSite intelligence, and applied it to other products in their portfolio to make the whole stack intelligent. They really did what they set out to do, which blows me away because typically they fall down on their face.

Nimble It's only available using iSCSI — Nimble can run iSCSI or Fiber Channel. Although iSCSI is cheaper, from a performance perspective, if you've got a really high-performance need, then you need Fiber Channel. 

Not to get too technical, but people tend to think that adding more ports gives you that much more bandwidth — that's not true. Every stream must be dedicated to a port. You could actually do port saturation and not use some of your ports that you just dedicated to your storage.

It's not as mature as Fiber Channel — no one is. Fiber Channel is more expensive, but if high-performance is what you seek, then you need Fiber Channel. I hope that they bring that forward. They're doing all kinds of leaps and bounds with the new Nimble. I really love where they're going, but I think Fiber Channel should be on the docket for dHCI. 

It would be nice if they just provided a simple, easy to read matrix (white paper) on the different models. Today, it's a little convoluted. It's a little hard to read. You have to actually show it to a customer and explain it to them. I shouldn't have to explain it to them. It should be very simple: here are your options, here's your expansion, here's how much storage you get with that, etc. The information is kind of there, but it's just, it's ugly.

I don't mean to call anyone's baby ugly, but they need to improve that datasheet. That would be my pet peeve with them. Their documentation needs to be simpler to read yet more detailed so it's easy to compare models and easier for clients to pick. They shouldn't have to have us sit down with them and explain to them why. They should be able to look at this and say, "Oh, okay. I understand where I need to be. I don't need a decoder to figure it out." That and the fiber channel are my pet peeves.

The documentation needs to be improved — the quick specs are just horrible.

I think there could be more automation capabilities; however, I think they're moving in this direction. More automation to open and close tickets, to get those service tickets open. They've done a lot in that area and I think that they again should focus on introducing more somewhere down the road. 

I have been using Nimble for roughly four years.

The stability is actually incredible because as they improve the older systems, the new operating system and new firmware deployed on it. It's now at six nines. So it's actually more reliable than it was when we bought it — it's fantastic.

I've never had an issue with Nimble Storage. The only problem that I've ever seen wasn't actually to do with the storage, It was an issue with the network. They didn't have their iSCSI set up correctly. It wasn't efficient. They weren't getting the throughput that they wanted. That was a simple change, but it had nothing to do with the storage; however, if you were to sit there and look at it with typical software that comes with VMware, it's going to say there is a storage latency issue. This is because it's looking at the time it takes to travel from the host to the storage and back. What it fails to see is that there are several components along that line. It doesn't mean it's your storage. InfoSite was able to see that, tell us, and show us where the real issue lied. The storage was acting very quickly, but the network was not.

It's very easy to scale. It's very simple to scale. Regarding the dHCI, if you put a new node on the network, it discovers it and says: "You want to add it?" — it's fantastic. I think they've done a real good job.

I didn't use them. I reached out to one of my friends at one of the distributors, because I know she knows the dHCI code — I talked to her about it. We have our own lab and we put one up in our lab and we started playing with it and picking it apart. It was pretty much straightforward, so I didn't have to call on them. For this reason, I can't really evaluate them.

I deployed quite a lot of different hyper-convergent products. 

The reason why I switched really came down to the intelligence, the InfoSite, and the fact that I really liked the idea that I can scale my storage independently.

A lot of people bought hyper-converged and then they said, "Oh, I need some storage. So I'll just use the storage on the hyper-converged and present that out to another server." That tends to ruin HCI. It will ruin your HCI, your ratios by doing that. You'll outgrow the box. You'll be all frustrated. And they were.

With Nimble dHCI, once I deploy it and set it up for dHCI, I can then take a portion of that storage and say, "Hey, you know what? I want to attach a sequel server to this." Then I am ready to start using this additional storage. I can sweat out my other assets.

Everyone always thinks greenfield is the way to go. Sure, of course it is. But nobody can afford a greenfield data center. That's just crazy talk. When I ran data centers for years, we would budget-cycle the data center. We'd say, "Okay, I'm going to compute on year one; on year two, I am going to do networking; and on year three, I'm going to do storage." They never are aligned with each other — the age of the equipment is never aligned. 

The refresh dates are never aligned. That makes it very difficult for those HCI stories. It's got to be the perfect storm. It really comes down to the fact that I need a better way to manage the infrastructure. That's when it becomes a more viable story. People say to me, "Well, I can buy servers cheaper here than I can in Microsoft Azure." Of course you can, but here, you're comparing a banana to an orange — they're not the same. The orange is obviously more tasteful.

I don't like bananas but If that's your thought process, that: "Hey, I want to refresh servers and that's all I want to do." Well, okay. Go ahead. Go ahead and do that. But you're constantly in that cycle. Why do you spend so many cycles refreshing? Why do you spend so many cycles maintaining infrastructure? Why is it that you have so many different tools to manage your infrastructure? The answer is because you won't get out of that mindset.

You have to step back from that mindset and say, "Imagine If I didn't have a data center. How would I do it today if I had a greenfield?" It doesn't mean you have to go out and start a greenfield — you start a little section that's green, move through your data center as you cycle, repurpose your budget to build the new infrastructure in a new way that's easier for you to manage and to procure.

A customer could set this. A customer with zero experience could set this up. Just pick up the manual and go set it up. I happen to charge for that, so don't tell them I said that.

I think the price is good. 

I hate talking about products. I think we should sit down and talk about what you are trying to do? What are your business goals? Because your business dictates to you what projects are important to them. What are they trying to accomplish? How do I help you accomplish those goals?

A lot of times, what it comes down to is they don't have the time to do the infrastructure the way we used to do infrastructure. That's no longer acceptable by the business because it's not important to the business. Think about how you can be a good business partner and provide them with the applications and the data they need access to from anywhere in the world. Look at us now. We're all home. I'm in my basement. Everyone feels bad for me. My basement is a fully furnished basement with a pool table, a beer fridge, a 60-inch flat-screen TV. There's even a music section over here with a full drum kit and guitars. So don't feel bad for me.

At the end of the day, you're not in the IT business. You're in the business of making some products, whether it's t-shirts, shoes, chairs — I don't care what it is. Whatever your products are, that's your business. Your business is not IT. But IT is the tool that helps you sell.

What the business wants to know is, "Hey, I'm giving you a lot of money for IT? How did you help me sell that chair? I need to sell more of those chairs in order to pay for that IT equipment." How do you help them do that? One of the ways is to reduce the work that you have to put in. You can focus on the applications and the innovation that helps them sell that chair. We should be talking about how do we help the business become more agile. You should be focused on, "Hey, they're working from home, how can I get them better access?" Maybe you shouldn't be doing anything in the data center. Maybe you should be moving to that hybrid infrastructure.

dHCI with Nimble Will give you the ability to start growing your data into the cloud in a bi-directional way. Now you can start talking about technology once you understand what their problems are; however, you can't just go in and talk about technology. You have to talk about what the problems are. How are you going to sell? If their problem is they're trying to put a screw in the wall and you hand them a hammer, you're not really helping them out — you'll get the job done, but it's going to be ugly. 

On a scale from one to ten, I would give this solution a rating of nine.

There's always room for improvement. I think that more automation is needed. Certainly, they're getting there. I think that the reason why I am giving them a nine is that it's probably the best dHCI solution out there. From a storage perspective, outside of dHCI, I think it's one of the best storage solutions out there in the market. I think it's priced right and it's extremely intelligent.

My biggest pet peeve is you need to give IT folks time back. The worst thing in the world is when you've got a problem with a system and you call support and you get level one and they go, "Okay, I need you to go in and I need you to dump these reports for me." The brilliant thing with Nimble is you don't have to do that.

They already have the reports because InfoSight provides them with the data all the time. So you don't call level one, you don't call level two, you go directly to level three. To me, that's the way it should be. 

You've got the data. Doesn't your team know how to read it? That to me is the most important feature of Nimble and what puts them on top of my charts. Because it's truly intelligent. Not only that, they're taking advantage of the intelligence of it to actually be proactive for their customers.

Our usage of Nutanix Acropolis AOS focuses on normal workloads for OSS-related applications of our operations support systems, e.g. customer facing systems, DNS, DHCP, and DPEs. The word we mention to our customers that describes this technology is OSS (Open source  software). This is not 5G.

One of the good features of Nutanix Acropolis AOS is the Prism Central. It's a well-designed system. It's a central installation and management tool for more than one cluster. If you have several clusters crossing Europe, you'll have a very fancy, smart, and tiny tool to orchestrate this.

Another feature I find valuable in Nutanix Acropolis AOS is technical, e.g. the rebuild time from the resynchronization between the clusters is fast. If you have an outage caused by a disk failure or a server failure, the system works well and is very fast, because the technology of Nutanix Acropolis AOS is even better than other systems. The system works faster than others.

Whether this system is worth the money will really depend on your goal, on what you want to achieve. The flexibility of Nutanix is very, very good. You have the freedom to use any kind of major vendor, to use it for a hyperconverged installation.

You have the Acropolis hypervisor based on KVM, e.g. the Kubernetes virtual machine. You can also use OpenStack. The system also supports VMware and Microsoft Azure. It also has integration with a hybrid cloud, e.g. AWS. There's also a good integration between the system and Google GCP Cloud. Nutanix Acropolis AOS has finesse. It's very, very good.

The look and feel of the web GUI of Nutanix Acropolis AOS needs improvement, when compared to other systems, e.g. VMware Orchestrator. For example, finding important features of the system should be easier. The features should be made more visible and easier to find, rather than having to figure them out and reconfigure them.

Another area from improvement for the system is hardware integration. I had some issues with the integration with the hardware vendor, in particular, Dell. The integration was really tricky, but the reason could be between the two vendors: Nutanix and Dell, because they have different life cycles for the deployment. The integration issue could be because of the hardware, firmware of Dell, and Dell had a different life cycle for the renewal of an update of the firmware, for the servers like Nutanix. We also had some issues with some Dell-related drivers, and that consumed a lot of time.

A one-touch system for integration could be an improvement. Having a one-touch update is also a very good idea. For example, you'll just need to push the button for the system to be updated automatically, e.g. for updating the firmware, hardware, disk, etc.

Nutanix Acropolis AOS which was integrated with Dell was not running well. What was running well was less than 50%, and the other updates have failed because of issues between the firmware, the server, and the system. This is why hardware integration with this system needs improvement.

I started using Nutanix Acropolis AOS in 2017, so the total number of years I've been using it is five.

I find Nutanix Acropolis AOS stable. In the beginning, we had some issues with its stability, but now, stability for it is good.

The system is very scalable. It's similar to vSAN, but vSAN is a well-known, major vendor for virtualization, while Nutanix Acropolis AOS is better in terms of technical features.

Technical support for Nutanix Acropolis AOS was very good. Their engineers have very good skills, e.g. in every case, they will try to solve an issue.

Technical support from Nutanix was better than VMware technical support. VMware was also very good, but the Nutanix support team had more initiative. Nutanix takes more care of the open tickets. I'm giving them a thumbs up as they're good.

I was involved in the deployment process for Nutanix Acropolis AOS. The system is working okay. It's good. The deployment the first time was a bit tricky, but we had good support from our vendor. I was able to run the installation by myself, with some support. You have to do some preparation for the installation to be good.

You have two Nutanix Acropolis AOS license models. One model is coupled with the hardware appliance, e.g. it's the appliance approach, depending on the vendor: HP, Dell, and Supermicro carry these appliance permit licenses. The other model is the software only model.

We had the first model that was coupled with the hardware from a vendor. For Nutanix Acropolis AOS itself, the pricing was okay, especially because of its features and how you can use it for a lot of hypervisor technologies. The cost, however, when coupled with the hardware vendor was high.

On our project, the renewal was from the hardware vendor and it was expensive. The software was okay, e.g. it had normal pricing.

You can choose to pay the license yearly, or you can pay every three years, or every five years. It depends on how deep your pocket is.

I evaluated VMware vSAN.

Nutanix Acropolis AOS is deployed on hardware, without a layer in between. It's server hardware installed.

Nutanix Acropolis AOS was good, e.g. from its performance running faster than VMware ESXi on a vSAN, so I'm rating it an eight out of ten.

In medium-sized installations, many of our customers require a reasonably-priced solution to replace legacy computer storage. Those are the customers who we are suggesting using the  VxRail solution to. If the customer is building a data center using VMware, we are recommending VxRail because of the VMware compatibility.  

Medium set-ups can be used in a large enterprise but only when they need this solution as a smaller part of their environment. It works well for small setups or medium setups, or for new application setups.  

Most of the products of this type have features that are almost the same. We are using VxRail especially because we would like to have products that are compatible with VMware and Dell to support our prospective client base.  

The configuration of HCI (Hyper-Convergence Infrastructure) solutions is very easy compared to the legacy solutions. Legacy systems run the computer and the storage separately and use switches to get the connectivity. That is much more complicated. It is completely the opposite when using generic HCI technology. The implementation is very simple and so is the operation.  

The only thing about the HCI solution is that troubleshooting is a little bit difficult because it is still a new technology. Other than this it is simpler than the traditional technology. HCI is nice and it makes sense. I think there is a need to improve the solution because it is difficult to troubleshoot. But compared to legacy solutions, you are troubleshooting one that is a little bit difficult instead of troubleshooting two different products that might each be a little easier. In the end, the difference as far as troubleshooting is not much but the advantages are still there on the HCI side and technology upgrade.  

The other thing I would like to see improved is not really a feature. It is about scalability. It would be good to increase the limit of the number of nodes within the clusters.  

I had experience before 2018, but I have been using it for this past year.  

Even compared to Cisco, I think VxRail is a very stable solution. It is in the same class as Cisco.  

VxRail is easy to scale, but the number of nodes is limited. This is not the same with other solutions like Nutanix. VxRail has a limit for a certain number of nodes within the cluster and if you need more than that then you have to create another cluster. It is an issue but at the same time, it is not an issue. It is kind of just a configuration difference.  

I do not have any direct contact with technical support because I am not doing the product delivery. If the technicians have some issues, they have to make the contact. I have not heard anything bad or good about the support. That suggests it is good.  

The initial setup was simple. The amount of time it takes depends on the number of nodes you have.  

It does require some maintenance over time. For maintenance, you have to request that through your supplier or even through the implementation team. It will be totally different depending on the kind of activity and the issue, but it should not be disruptive for the most part. The only exception is in critical applications. These may be critical but it is simple to restore the network connectivity or storage availability.  

My advice about cost and setup is that it is just as cheap to have the HCI solution as to stay with legacy solutions. If you convert the value of HCI versus the traditional, legacy solution you gain more than you spend. It comes out to an even trade as far as budget.  

Compared to other HCI products, VxRail has advantages in stability and support from the vendors. If you bought Cisco, for example, you have to open an account with Cisco for the hardware and open an account with VMware for the software separately. With VxRail this is not what happens. You opened the account with them and they will manage all the communication and the services. That ends up being more stable. Getting your support from one company is better than having to get support from different companies when dealing with an issue shared between products and trying to sort that out.  

I recommend VxRail as a solution especially for those using legacy services. We often recommend VxRail over other competitors. The only exception really is if the customer does not want Dell computing resources. For example, the client may have another vendor they tend to use. So if everything they have is HP they may like to have an HP solution. If they are using Cisco, they may want a Cisco solution. Those are the only times that we will not go with suggesting VXRail.  

On a scale from one to ten where one is the worst and ten is the best, I would rate the VXRail solution as an eight-out-of-ten.  

We use Sangfor HCI for the Sangfor Cloud platform. We are soon going to host our own cloud platform service by using Sangfor HCI. We've already got a conditional queue from one of the customers, and we're consistently in talk with Sangfor regarding this deal.

We are working with the trial license of the Sangfor platform. For the first three months, we're going to use it on a trial basis, and after the customer confirms the purchase order, we'll be permanently going for the Sangfor Cloud solution.

It should have object storage. It already has network virtualization and micro-segmentation, but it is missing object storage. It does not have object storage. vSAN also does not have object storage, but Nutanix has object storage, and it is natively integrated into its HCI. So, if you pay for the object storage license, you have that module pre-built. 

I would've been really happy if we had point-to-point network connectivity between the virtual firewall and the customer site lease line so that we could have a private IP VPN.

Sangfor is relatively new in all markets. I would request the Sangfor team to come up with a partner training program with a free voucher so that we can get some of our staff trained and certified in Sangfor and educate them in a proper way.

It has been about five months since we've been using the Sangfor platform.

We've been using it only for about five months, and I will be able to comment on the stability of this platform after about two years. I've been working with VMware for 10 years. So, there is no point in comparing it with VMware.

We're going with two nodes at the moment. There will be some downtime, but we can manage it for the next two to three months because we have fewer customers. We're going to provide the cloud service on the Sangfor platform, and soon, we'll be upgrading to the third node. After we have three nodes, it's going to be easier for us to scale up the platform.

We are a team of about 25 individuals. Four people are looking after this solution. When we start to scale up, we would need more people. With the license that we will purchase, we will also get support from Sangfor. If there are any specific cases, we will contact Sangfor support.

Previously, we had Harper from their team for support. He has now moved to a different role, and he was very good. I would rate him 4.5 out of five. 

Their technical support is pretty good, but we are having some language issues. Basically, we get support from China, and we are having some issues with English. I would rate them a four out of five.

We've worked with VxRail and vSAN in the past. We've not used Nutanix, but I would like Sangfor to grow as a competitor that could beat Nutanix. That's because the RFPs that are made are so much based on Nutanix. It is really difficult for VMware and vSAN to beat Nutanix. I would like to see Sangfor grow as a company and beat Nutanix with all its features. If they add object storage, it would be very good.

For me, it was initially a bit complex, but later on, I gave it to my team. They communicated with Harper and other support staff from the Sangfor team, and they were able to do it. It is not a difficult task for us now. We've already given the Sangfor virtualization platform to a customer. So, we already have two Sangfor clients, and we are looking forward to increasing the number.

Sangfor needs to be more aggressive because this is a new market or territory for Sangfor. Nepal is a very price-sensitive market, so Sangfor needs to be a little more aggressive with its pricing. I would rate them 3.75 out of five in terms of the price.

We're comparing it with VMware. 

To anyone looking to implement Sangfor, I would advise doing a three-month PoC. We have a system deployed right here in our test labs with two nodes, and we can give a Sangfor cloud platform tenant for a PoC. We are always ready to support customers, and I believe Sangfor is too.

It has only been about five months since we have been using it, and I'm pretty satisfied with it. I would rate it an eight out of 10.

We're using SANsymphnony for our primary storage in a HA environment for sensitive production data. Storage nodes are HPE servers with SSD drives in them. They serve storage to two servers blades enclosures in a Fibre Channel storage network. We have a total of 50TB in a mirror. 16 servers use this storage in a VMware vSphere environment.

This infrastructure has run without any issues since 2017 and we update it twice a year.

Before that, we had other SANsymphony infrastructure running on an old HMP MSA storage array. We migrated without any interruption.

DataCore SANsymphony brings the ability to work on almost every hardware platform and to build a storage node with high precision (disks, interfaces, protocols). We can choose what hardware we want to put in and it brings a high-performance throughput from it. The ability to build exactly what you need is a major advantage of SANsymphony over other solutions.

Hardware maintenance is easy since it is a simple server. There's no need for a storage hardware expert. On top of that, the Software-Defined Storage is easy to manage.

We used the wizard to deploy SANsymphony in a virtual environment for hyper-converged infrastructure and it is quite useful. We can deploy virtual high-available infrastructure for running tests in less than an hour.

Continuous data protection is a powerful feature that can save your data in several cases. It is good protection against ransomware.

The ability to use the RAM of the server node as a read and write cache brings a lot of performance to the storage. We can reach high IOPS from slow disks through the huge amount of cache.

The documentation repository is really useful and kept updated.

We would like to see a real "sexy" storage dashboard with capacity, usage, performance, and error tracking.

The cloud reporting interface is quite poor compared to other vendors. We are far from an HPE Infosight, for example.

Using a classic storage array constructor allows clients to have a single point of contact in case of an issue. With DataCore, we have to deal with them for the software part and with the hardware vendor for the hardware part. Sometimes, in a complex environment, we have to deal with storage array vendors, servers vendors, and software vendors and that can be exhausting.

We've been using SANsymphony for almost 10 years now. We followed all of its great evolution so far.

This solution is stable. Depending on the two rooms' interconnection, we have to deal with redundancy and maybe a witness. Losing the connection completely between the nodes can lead to a complete rebuild of a side.

The scalability is almost infinite. I cannot think of a more scalable solution.

The customer service is quick and efficient. Sometimes we have to deal with some trivial questions but that seems to be the basis of every support query right now.

We previously used an HPE 3PAR storage array. We switched to have more performance and more flexibility. The maintenance is quite easier too.

The initial setup is easy via the use of a comprehensible wizard.

We implemented the solution ourselves after a formation from the vendor.

The ROI is high as we can change or update hardware without changing our licencing. 

The main work in building a SANsymphony solution is to design and select the correct hardware parts. The setup is quite easy and the configuration is too.

The licensing is by terabyte and can be quite expensive. 

We evaluated other hardware vendors but stuck to a software-defined solution.

VMware VSAN was not mature enough for us and we did not want to use a hardware vendor solution.