Comprehensive software reviews to make better IT decisions
Hyperconvergence for the SMB
Questions about hyperconverged infrastructure HCI from organizations with small IT footprints are a good example of the difference between an SME and an SMB when it comes to product selection. Though the “S” in both cases refers to “Small” the differences between “E” and “B” can be significant.
Hyperconvegence has been getting a lot of traction in the Small to Midsized Enterprise or SME space, particularly for companies moving off legacy external storage (SAN or NAS array) and looking for more streamlined operations. Smaller shops with more modest virtual machine installation – let’s say under 30 VMs – are finding hyperconvergence products cost prohibitive.
Here is a typical scenario:
- An IT department with about 15 or so servers got into virtualization a decade ago to reduce the cost of server replacement, consolidating the 15 workloads onto two or three servers using VMware virtualization.
- For availability and resiliency they also bought a modest external storage array shared between the VMware host servers.
- The servers and/or the storage has reached end of life. Hyperconvergence is getting a lot of press. Is there an HCI option for this situation?
- Upon investigation it is found that solutions from the big names in HCI, like Nutanix, are just too expensive and too “big” for the small shop.
- But what about VMware vSAN software running on commodity server hardware? This option is also deemed cost prohibitive due to the licensing costs and the number of “nodes” required – 3 but ideally 4.
So are there any credible HCI options for the smaller and more budget-conscious shops? Yes, there are. But you should also give some thought to whether you even need an HCI solution.
Scale Computing and StarWind for the SMB
In the case of HCI for SMBs there two vendors that come up regularly among Info-Tech’s clients: Scale Computing and StarWind. These are not the only options, but they have both been getting a lot of SMB buzz recently.
In terms of technical features, the two have a lot of similarities and their HCI feature sets are similar. One differentiator has to do with how much the customer needs or wants to continue to use VMware for virtualization. A lot of SMBs that are VMware shops have first looked at VMware vSAN HCI nodes but baulked at the cost.
- StarWind is “hypervisor agnostic.” That is to say it uses Microsoft Storage Spaces to virtualize the storage across nodes and can host a range of hypervisors including Microsoft Hyper-V and VMware vSphere. So if you want to park your VMware VMs on a HCI without vSAN then StarWind is the better bet.
- Scale Computing’s HyperCore platform uses open source code for storage virtualization and KVM for its hypervisor. So migrating to Scale would involve a conversion of your current VMs to KVM. But Scale does offer support for such migrations. Not surprisingly, Scale is the darling of the happy to be Ex-VMware crowd.
But, Do You Even Need Hyperconvergence?
Another option to consider is whether you even need an HCI node cluster. Consider what is driving the hyperconvergence trend. The three top drivers in order of importance are:
- Simplification of management. The all in one approach removes the need for separate management regimes (and managers) for servers, network, and storage.
- Ease of scaling. HCI has scale out capacity. Need more processing and storage? Add another node to the cluster. No rip and replace or data migration headaches.
- Reduce hardware costs. HCI can reduce the overall footprint and costs of infrastructure. But that does not mean HCI is low cost.
For an SMB with modest and stable VM implementation, it is arguable whether reason 1 and 2 even apply to their situation. Management is likely already amalgamated in one person/role and the infrastructure is not likely to need major scaling. As for reason 3, we’ve already established that many HCI offerings are too big and expansive.
Ten years ago even small businesses were looking at host clusters and an external array for their VM infrastructure. Servers just did not have the capacity – processing and storage – to host everything in one box. External storage also enabled availability features (such as host-to-host VM live migration and high-availability failover).
But today we have servers that are orders of magnitude more powerful and have greater storage capacity. A small shop can conceivably host everything on one server host with direct attached storage.
Isn’t that putting all your eggs in one basket? What about availability and fast recovery if something should go wrong with the host? This is where software-based replication can come in. Here you can have two hosts – one primary and the other a passive alternate – and do replication between the hosts.
- Focus on your storage requirements. Hyperconvergence brings previously externalized storage back to the virtual server host nodes. Leaving aside HCI for a moment, list your storage requirements. Then look at whether those required features are supported in HCI.
- Consider the options beyond HCI. Chances are your requirements will be met by various HCI options. But they also may be met by alternative setups. For example, if the main requirement of your storage is support for multi-node resilience, you may be able to achieve that with software-based replication as outlined above.
- Consider Scale and StarWind among SMB options. These are not the only options for the SMB, but they are the two we hear about the most often. Neither was included in our most recent HCI vendor landscape, but this does not mean they are not worth considering. StarWind just didn’t have the profile when we put together our initial vendor list. That has changed over the past year. Scale asked us not to be included. Normally we don’t let vendors “opt out” of one of our landscapes, but Scale made the case that they were not a good fit for our evaluation criteria which was (unintentionally) weighted against its product.
Organizations with smaller, more modest, server infrastructures looking to refresh storage and servers with hypeconvergence should consider if they even need HCI. While there are lower-cost HCI options targeted at SMBs worth considering, be sure to also consider if there are other options – such as host servers with software-based replications – that will meet your requirements.
Want to Know More?
COVID-19 has forced software companies and their suppliers to refocus efforts around prioritizing systems and workflows that are nearly 100% digital in nature. As a result, Info-Tech has observed the quick emergence of six market themes that are highly relevant after COVID-19. This note series will profile key vendors and how they fit into the post-COVID-19 world.
Oracle has announced the general availability of Exadata Cloud@Customer, a managed service that enables enterprises to unlock the previously cloud-first features of Oracle's Autonomous Database for on-premises data centers. This offering is ideal for enterprises that must conform with regulatory and/or technical challenges that force on-premises database residency.
More than ever, cybersecurity solutions are core to any MSPs offering. No longer should technology service providers be farming this out to dedicated security providers. Trust and peace of mind are the core tenets of what they are selling and solutions like Acronis Cyber Protect Cloud can provide the platform upon which to deliver on those promises.
Microsoft Cloud Services Usage Surges 775% for Teams in Regions With Enforced Social Distancing – Part 2
Experiencing issues when using Microsoft online services? You are not alone. Capacity constraints were being hit, pre-COVID-19, and usage has surged in regions with enforced social distancing.
Google has announced a premium support plan for its cloud customers, promising a 15-minute response to the highest severity tickets. Google’s cloud has long struggled with enterprise customers – especially when compared to giants Microsoft and AWS – and this announcement is the latest incarnation of Google’s push to better serve a critical constituency.
Microsoft Announces Expansion of Azure Canadian Infrastructure, Offers Data Residency and High Availability
In January, Microsoft announced what it’s calling “the largest expansion of its Canadian-based cloud computing infrastructure” since 2016. Additional availability zones and services will increase capacity for cloud-hungry Canadians, and the addition of an Azure ExpressRoute site in Vancouver will guarantee security and performance in a regulated jurisdiction.
Microsoft’s announcement that server-side encryption with customer managed keys for Azure Managed Disks is now available is welcome news for security-minded public cloud customers. Managing one’s own keys in a cloud environment can be an important step in complying with regulatory requirements, and this new feature should open Azure Managed Disks to a wider group of customers who may have held back for this reason.
Amazon Web Services (AWS) has provided its customers with better options for Virtual Private Cloud (VPC) ingress routing. Customers will have to consider which works best for their needs.
AWS VPC Traffic Mirroring gives customers more visibility for out-of-band traffic inspection. This feature is another useful tool for monitoring in the AWS cloud.