Hyper-Converged (HCI) Forum

Michael_Davis
User at Lennox International
Apr 05 2021

Hello,

My environment is as follows:

  • APP server
  • Web servers
  • SQL servers
  • File servers

We have 25 servers and 100 users. 

Use case: Production and Distribution

Which Hyper-Converged Infrastructure (HCI), in your opinion, is the best and the most suitable (for my environment) for a disaster recovery solution?

Steffen HornungSo you have 25 physical servers you want to virtualize, or are there 25 VMs which need a DR solution? In both cases: this is a rather small setup. We host over 20 VMs per physical Host.  You could virtualize with ESXi and protect the VMs with Veeam. 3 Hosts for virtualization and 1 or 2 Hosts for disaster coverage in case that disaster strikes. But that would have to be on another datacenter (if you have the space or a 2nd location, cloud would work also). Not really HCI, but works. With Nutanix you could use a single block (3 or 4 nodes in single chassis) for production and maybe a small 2-node-cluster to use for disaster recovery. With Nutanix Xi Leap you can leverage cloud if a dedicated DR cluster is too much. With Nutanix you have free hypervisor choice as they can use ESXi, Hyper-V and their own AHV. ESXi would contribute to more license costs of course. Cloud depends on your internet uplink. So be sure to look into that when considering. My personal preference is Nutanix. Their support is just awesome.
Rony_Sklar
IT Central Station
Apr 05 2021

What are some important factors to keep in mind and to compare when choosing between HCI solutions? 

ShivendraJha1. Support 2. Migration or Conversion process from existing solution 3. Cost  4. Hardware compatibility  5. Integration with all critical and non-critical solutions 6. Cloud readiness
Gaurav VyasAvailability, support, cost, compatibility and scalability, cloud readiness.
Michael SamaniegoThere are several solutions that claim to be HCI in the market, however the best factor is the native integration with the hypervisor without the need to have additional virtual machines that "perform HCI", so far in several cost-efficient scenarios that I have performed and in turn With different hardware manufacturers I can personally say that the best option is VMware vSAN. Its main strength is the correct management of hardware resources.
Ariel Lindenfeld
Sr. Director of Community
IT Central Station
Dec 09 2020

There are a lot of vendors offering HCI solutions. What is the #1 most important criteria to look for when evaluating solutions?

Help your peers cut through the vendor hype and make the best decision.

Bharat BediWhile there is a long list features/functions that we can look at for HCI -> In my experience of creating HCI solutions and selling it to multiple customers, here are some of the key things I have experienced most customers boil it down to: 1) Shrink the data center: This is one of the key "Customer Pitch" that all the big giants have for you, "We will help you reduce the carbon footprint with Hyperconverged Infrastructure". It will be good to understand how much reduction they are helping you with. Can 10 racks come down to two, less or more? With many reduction technologies included and Compute + Storage residing in those nodes, what I mentioned above is possible, especially if you are sitting on a legacy infrastructure. 2) Ease of running it: The other point of running and buying HCI is "Set it and forget it". Not only should you look at how easy it is for you to set up and install the system, but how long does it take to provision new VMs/Storage, etc. It is great to probe your vendors around to find out what they do about QOS, centralized policy management, etc. Remember that most HCI companies portfolios differ at the software layer and some of the features I mentioned above are bundled in their code and work differently with different vendors. 3) Performance: This could be an architecture level difference. In the race of shrinking the hardware footprint down, you could face performance glitch. Here is an example: When you switch on de-duplication and compression, how much effect does it have on the overall performance on CPU, and thereby affecting the VMs. Ask your vendors how they deal with it. I know some of them out there offload such operations to a separate accelerator card 4) Scaling up + Scaling out: How easy it is to add nodes, both for compute and storage? How long does it take while adding nodes and is there a disruption in service? What technologies do the vendors use to create a multi-site cluster? Keep in mind if the cluster is created with remote sites too? Can you add "Storage only" or "Compute only" nodes if needed? All of the above have cost implications in a longer run 5) No finger pointing: Remember point number two? Most of these HCI are based on "Other Vendors' hardware" wrapping it with their own HCI Software and making it behave in a specific way. If something goes wrong, is your vendor okay to take full accountability and not ask you to speak with a hardware vendor? It will be a good idea to look for a vendor with a bigger customer base (not just for HCI but compute and storage in general) - making them a single point of contact and more resources to help you with, in case anything goes wrong.
SamuelMcKoyIn my opinion, the most important criteria when assessing HCI solutions other than the obvious performance. How does that HCI solution scale? Or in other words, how does one add storage and compute resources to the solution. Without understanding how the solution scales one can easily request resources without understanding how and why the overall costs have ballooned. The costs can balloon not only because you're adding additional nodes to your HCI cluster for the additional storage and compute resources that were needed but also with additional compute nodes added to the cluster this requires additional licensing for whichever hypervisor the HCI solution depends upon. This is usually on a per-compute-node basis. For example, some HCI architecture allows admins to add only storage to the HCI cluster when additional storage is needed. Not requiring the purchase of any additional licensing from the hypervisor's perspective. On the other hand, some HCI architecture requires you to add a compute node with the additional storage you need. Even if you don't need the compute resources required to add that storage. That compute node will then need to be properly licensed as well. This type of architecture can and usually does force its consumers to spend more money than the circumstances initially dictated. So for me how the HCI solution scales is most important because it can ultimately determine how cost-effective the HCI solution really is.
Bart HeungensFor me an HCI solution should provide me: - ease of management, 1 console does all, no experts needed, cloud Experience but with on-premise guarantees - invisible IT, don't care about the underlying hardware, 1 stack - built-in intelligence based on AI for monitoring and configuration - guaranteed performance for any workloads, also when failures occur - data efficiency with always-on dedupe and compression - data protection including backup and restore - scalability, ease of adding resources independent of each other (scale up & out) - a single line of support
Rony_Sklar
IT Central Station

How does hyper-converged differ from converged? 

Is one better than the other? When would one choose converged, rather than hyper-converged?

Are there pros and cons to each type of solution?

Dan ReynoldsHyper-converged is typically an "all in one box/rack" solution. It consists of compute, storage & network resources all tied together physically (and through software).  Hyper-converged for a pro - is a complete solution. You don't have to architect it. All you have to know is how much "power" you need (what you want to do with it). While with converged infrastructure (which can still be 'software defined') you have to match and configure the components to work together.  More often then not converged infrastructure is cheaper. You might already have the storage and networking resources, for example. And manufacturers put a premium on packaging the solution together. 
PierreChapusHyperconverged is a system cluster of at minimum 3 nodes. The system mirrors datas between nodes and runs virtual machines.  Converged systems is anything between the classic server and hyperconverged platform. This converged concept was useful in waiting for hyperconverged development and should disappear in a near future.
Satish Dgconverged infrastructure still incorporates hardware, running the technology natively on hardware. On the other hand, hype convergence is fully software-defined and completely integrated
Rony_Sklar
IT Central Station

From my own research, it seems that Converged Infrastructure relies on hardware, whereas Hyper-Converged Infrastructure is software-based. What does this mean in practical terms? What are the pros and cons of each?

ROBIN JACKSONIn principle you’re right “Converged Infrastructure relies on hardware, whereas Hyper-Converged Infrastructure is software-based”. But there are further advances for software management of containers, VMs, storage, and networks within a single architecture. As a Red Hat partner, we are aware of coming developments based on Red Hat OpenShift which significantly simplify operations and provide complete management and portability across On-Prem, Hybrid, and Multi-Cloud environments.
Bart HeungensAlso in a converged infrastructure software is important. Converged for me is a combination of hardware components that are sold as a single solution and where a software layer is added to make the management easier. But the hardware solution consists mostly from individual server, storage and networking components. Most hyperconverged solutions goes further with integrating the storage layer into the server layer, removing a layer of hardware, and where the software inside the solution create a shared storage pool for the server stack. Automatically the management layer is also simplified just as with the converged solution... Less hardware (or differently used) and more software inside... I call it more a typical evolution of IT infrastructure... Know that converged and hyperconverged is a marketing thing and not really a product as such... I saw converged and hyperconverged solutions already 20 years ago before it even existed... Just look for what you need and pick the right solution... 
Norman AllenA Converged Infrastructure has more hardware.  Compute is on one set of hardware.  Storage is on another set of direct-attached (or other) hardware.  Networking is separated, too. In a Hyper-Converged Infrastructure, Compute and Storage are on the same hardware, and depending on the complexity of the solution, sometimes Networking isn't even needed because you can directly connect the nodes to each other if you only have 2 nodes.  Adding nodes is as simple as duplicating the hardware and scaling up or out, accordingly.    A Hyper-Converged Infrastructure requires less hardware and gives you a more simplified solution.  It is also less expensive to procure, operate and maintain.   
Rony_Sklar
IT Central Station

What are key factors that businesses should take into consideration when choosing between traditional SAN and hyper-converged solutions?

reviewer1234203There are so many variables to consider. First of all, have in mind that tendency is not the rule, your needs should be the base of decision, so you don't have to choose HCI because it's the new kid on the block. To start, think with your pocket, SAN is high cost if you are starting the infrastructure; cables, switches, and HBAs are the components to add to this structure that have a higher cost than traditional LAN components, On the other side, SAN requires more experimented experts to manage the connections and issues, but SAN has particular benefits sharing storage and servers functions like you can have on same SAN disk and backup and use special backup software and functionalities to move data between different storage components without direct impact on servers traffic. SAN has some details to consider on cables like distance and speed, its critical the quality or purity to get the distance; the more distance, the less speed supported and transceiver cost can be the worst nightmare. But SAN have capabilities to connect storage boxes to hundreds of miles between them, LAN cables of HCI have 100 mts limit unless you consider a WAN to connect everything or repeaters or cascaded switches adding some risk element to this scenario. Think about required capacities, do you need TB or PB?, Some dozens of TB can be fine on HCI, But if there are PBs you think on SAN, what about availability?, several and common nodes doing replication around the world but fulfilling the rules of latency can be considered with HCI, but, if you need the highest availability, replicating and high amount of data choose a SAN. Speed, if it is a pain in the neck, LAN for HCI starts minimum at 10 Gb and can rise up to 100 Gb if you have the money, SAN has available just up to 32 Gb and your storage controller must be the same speed, this can drive the cost to the sky. Scalability, HCI can have dozens of nodes replicating and adding capacity, performance, and availability around the world. With SAN storage you can have a limited number of replications between storage boxes, depending on manufactures normally you can have almost 4 copies of the same volume distributed around the world and scalability goes up to controllers limits its a scale-up model. HCI is a scale-out model to grow. Functionalities, SAN storage can manage by hardware things like deduplication, compression, multiple kinds of traffic like files, blocks or objects, , on HCI just blocks and need extra hardware to accelerate some process like dedupe. HCI is a way to share storage on LAN and have dependencies like the hypervisor and software or hardware accelerators, SAN is the way to share storage to servers, it is like a VIP lounge, so there are exclusive server visitors to share the buffet and can share the performance of hundreds of hard drives to support the most critical response times.
Tim WilliamsWhether to go 3 Tier (aka SAN) or HCI boils down to asking yourself what matters the most to you: - Customization and tuning (SAN) - Simplicity and ease of management (HCI) - Single number to call support (HCI) - Opex vs Capex - Pay-as-you-grow (HCI)/scalability - Budget cycles If you are a company that only gets budget once every 4/5 years, and you can't get any capital expenditures for Storage/etc, pay-as-you-grow becomes less viable, and HCI is designed with that in mind. It doesn't rule out HCI, but it does reduce some of the value gained. Likewise, if you are on a budget cycle to replace storage and compute at different times, and have no means to repurpose them, HCI is a tougher sell to upper management. HCI requires you replace both at the same time, and sometimes budgets for capital don't work out. There are also some workloads that will work better on a 3Tier solution vs HCI and vice versa. HCI works very well for anything but VMs with very large storage footprints. One of the key aspects of HCI performance is local reads and writes, a workload that is a single large VM will require essentially 2 full HCI nodes to run, and will require more storage than compute. Video workloads come to mind for this. Bodycams for police, surveillance cameras for businesses/schools, graphic editing. Those workloads can't reduce well, and are better suited for a SAN with very few features such as an HPE MSA. HCI runs VDI exceptionally well, and nobody should ever do 3 Tier for VDI going forward. General server virtualization can realize the value of HCI, as it radically simplifies management. 3 Tier requires complex management and time, as you have to manage the storage, the storage fabric, and the hosts separately and with different toolsets. This also leads to support issues as you will frequently see the 3 vendor support teams blame each other. With HCI, you call a single number and they support everything. You can drastically reduce your opex with HCI by simplifiying support and management. If you're planning for growth up front, and cannot pay as you grow, 3 tier will probably be cheaper. HCI gives you the opportunity to not spend capital if you end up not meeting growth projections, and to grow past planned growth much easier as adding a node is much simpler than expanding storage/networking/compute independently. In general, it's best to start with HCI and work to disqualify it rather than the other way around.
Bart HeungensAll depends of how you understand and use HCI: If you see HCI as an integrated solution where storage is integrated into servers, and software-defined storage is used to create a shared pool of storage across compute nodes, performance will be the game changer of choosing for HCI or traditional SAN. The HCI solution of most vendors will be writing data 2 or 3 times for redundancy across compute nodes, and so where there is a performance impact on the applications due to the latency of the network between the nodes. Putting 25Gb networks, as some vendors recommend, is not always a solution since it is npt the bandwidth nut the latency of the network that defines the performance. Low latency application requirements might push customers to traditional SAN in this case. If you use HCO for ease of management through a single pane of glass, I see many storage vendors delivering plugins to server and application software, eliminating the need of using the legacy SAN tools to create volumes and present them to the servers. Often it is possible to create a volume directly from within the hypervisor console and attach them to the hypervisor servers. So for this scenario, I don't see a reason choosing between the one or the other. Today there is a vendor (HPE) that is combining traditional SAN in an HCI solution calling it dHCI. It gives you a HCI user experience, the independent scalability of storage and compute, and the low latency often required. After a time I expect other vendors will follow the same path delivering these kinds of solutions as well.
Rony_Sklar
IT Central Station

What are the benefits of using cloud versus hyper-converged infrastructure? What should enterprises take into account when choosing between these storage options?

Carlos EtchartI think that the key points to consider are: security, performance, and CAPEX vs OPEX Security: Having HCI on-premise allows you to keep your current security policies. For some customers having sensitive data on the cloud is not even an option due to their policies. If you go to the cloud you must remember that you are responsible for the security of your data, not the cloud service provider and new policy schemes may be needed. Performance: You have to evaluate if the cloud provide the bandwidth, throughput, and availability that your operation requires vs. on-premise. CAPEX vs OPEX: Even though there are some schemes that allow you to have HCI on-premise as EaaS (Everything as a Service like HPE GreenLake) most of the customers own their HCI infrastructure so depending on your expenditure convenience you will favor one or the other.
Tim WilliamsHCI is on-prem, so it's simpler and easier to manage and integrate with applications and your network. Something like a Nutanix can give you a lot of functionality of the cloud without having to deal with the massive headache that is designing your network and applications to be able to utilize the cloud effectively (for Infrastructure). SaaS is a fantastic use of the cloud, but infrastructure-as-a-service hasn't matured in process or manageability yet to justify. It will always cost more to be in the cloud, and it will always be more difficult to get to it. The cloud is amazing if you use it right.
Chaan BeardThere are several benefits of both Cloud and HCI that can be leveraged to the advantage of the feature rich HCI stack user hybrid style. The first is that many applications have not been designed for the cloud and require an on premise stack that can save data in the cloud and offer the same simplicity as cloud operations. If you select an HCI vendor that supports all of the Hypervisors and all of the clouds you can make your applications leverage each technology to your best advantage and lower OPEX costs by up to 60% without rewriting your applications to be cloud friendly. You can also simplify the entire stack and enjoy 5 microsecond latency and not make storage API calls that leave the kernel and introduce even more latency while they access storage from SAN and NAS devices. You can also serve up applications using FRAME technology with this stack that allows you to deploy solutions for remote workers in minutes that are fully secure. AOS offers full encryption and FIPS level 140-2 and better security built into the HCI stack right out the box, no need to go bolting on complex Frankenstein solutions like NSX that require several residents with deep knowledge of 8 different VMware stacks to operate the whole enchilada which increases your OPEX costs dramatically. AOS based HCI eliminates separate SAN, NAS and Object store silo's, it also eliminates system security and server with virtualization silo's and condenses them into one stack, so simple 8 year old children can administer it in a few mouse clicks. Mature HCI based also offers BC/DR benefits that will allow you to use the cloud for what the cloud os good at, BC/DR. Mature HCI vendors also offer their entire HCI stack for AWS and Azure so that you can drag Virtual machines from on prem to the cloud seamlessly. The San Jose based HCI vendor that does this is 4 years ahead of its competition (Dell-EMC) who only work with one Hypervisor while they work with any Hypervisor and all the cloud stacks concurrently. Nutanix Acropolis Operating System is the wave of the future and it runs on any hardware on their HCL from any server vendor. The HCL list is long. It is also a cluster based architecture that can be expanded one node at a time and they have GPU nodes as well. Nutanix Software Defined Valhalla is here today, so advanced everyone will think you are with the gods!!