Tech Field Day VFD4 – Scale Computing – bringing hyperconvergence to the SMB
Scale Computing was the very last presenter at Virtualization Field Day 4 in Austin, Texas, yet they are the very first presenter that I’m going to blog about. Now this has nothing to do with the fact that they had the “last word” or that they are the most recent vendor in my memory, it’s more because Scale achieved something that seemed a bit out of the ordinary for the week – they took a room full of VMware people (and one Hyper-V guy, sorry Jeff) and proceeded to peak everyone’s interest in a product that is based off of KVM. Scale’s presentation was one of the best IMO, the delivery, the format, the technology, the people they had there – it all seemed to mesh together well and I was very impressed. You can see it for yourself here.
Know your market
All to often we see vendors try and play the all encompassing role, catering to enterprise yet stating they also have offerings for small and mid market businesses and/or catering to SMB yet marketing their product at an enterprise data center scale at the same time This may be an easy thing for a larger company to do, say Dell, but for smaller companies it becomes a challenge – they have fewer employees and resources to truly be an “all in one” provider. Scale fits neither of these scenarios – right from the get go Scale has positioned themselves as a provider to the small and mid markets, and throughout their lifespan and their VFD4 presentation they didn’t veer away from that. Scale knows who their (potential) customers are and has certainly developed a solution that caters to them. As they stated, there are only 500 Fortune 500 companies, yet there are 380,000 small businesses.
What does SMB need?
Typically in the SMB space we see a very slim, if any, IT department – think anywhere from 1 to 12 people in your IT department doing everything from fixing digital signage, deploying iPhones, maintaining Exchange environments, and developing web applications/sites. It’s these type of companies that don’t have the resources, both people or money, to really employ any subject matter experts – they don’t have a virtualization guy, they don’t have storage guy, they don’t have an active directory guy, nor do they have any money to send employee’s off on training to become an expert in these technologies. So when it comes to running 50 or so virtualized workloads there are definitely challenges. These companies need a solution that is first off simple, something that they can simply plug in and use without too much trouble or training. It also needs to be a solution that is cheap – more times than none SMB enters into the buying phase, especially in IT, without a budget. They don’t have a fixed amount to spend on IT so if they go to the CEO or CFO with a solution that is priced through the roof you can pretty much bet on being shot down. One more thing SMB may look for in a solution is scale – let’s face it, most SMBs don’t enter a market with hopes of simply being an SMB their entire life – they want to grow, they want to expand, and with that, their technology solutions will need to adapt.
What does Scale offer?
Although Scale Computing originally went to market as a scale out storage solution, their initial intent was to always be in the hyperconvergence space. As Jason Collier stated during his presentation, “Scale Computing, a pretty stupid name for a storage company right?” They’ve successfully done this by coming to market with their hyperconverged solution dubbed the HC3. The HC3 aims to be a “data center in a box” by lumping storage, compute and virtualization into a singular appliance offering a plug and play type deployment. HC3 is sold in three separate hardware platforms; the HC1000, HC2000, and HC4000, all of which are sold in a 3-node starter system taking up only 3U of rack space – perfect for cramming into a broom closet or the corner of the room like SMB so often does. Variations between the platforms are as follows; The HC1000 system – a SuperMicro SATA based solution ranging from 6-24TB raw storage, 96GB RAM, and 12 cores of CPU. The HC2000, a Dell based SAS solution, ranging from 7.2 – 14.4TB raw storage, 192 GB memory and 18 cores. The HC4000, a Dell based SAS solution ranging from 14.4 – 28TB, 384GB Memory and 36 cores of CPU.
Although the hardware specs are important the real value of the HC3 comes from the HyperCore software which is layered on each node within your HC3 cluster. HyperCore, is purpose built for the HC3 system, continuously monitoring VMs, hardware, and software in order to ensure that nodes can automatically respond to infrastructure events and failures, maintaining the level of high availability that we all need. For the VMware faithful you can think of this like vCenter, but integrated into every single node – no separate license, no separate management server.
Underneath all of the orchestration included in the HyperCore software is perhaps some of the most innovative IP within Scale Computing and that is how they handle manage underlying storage. They do utilizing a technology that they have built from scratch, the Scale Computing Reliable Independent Block Engine, or SCRIBE for short. SCRIBE in layman’s terms is a direct block engine which bypasses any Linux file-systems whatsoever. It spans all of the HC3’s storage within the cluster, providing no single points of failure and ensuring consistency throughout the VMs running atop it. SCRIBE is what allows the HC3 to make it’s clones in a matter of seconds, as it doesn’t necessarily need to duplicate all data all the time. Same goes for snapshots and replication as well. For the VMware informed you could think of it a lot like VSAN, providing multiple copies of data spanning across multiple nodes without the use of a VSA. SCRIBE is a technical beast (at least for me) and how it updates metadata and ensures consistency can be a bit of challenge to understand for those that aren’t real storage nerds. I encourage you all to check out Scale’s explanation of SCRIBE from Storage Field Day 5 as it does a good job at laying it all out (get it ).
How does all this fit into SMB?
Sure when you pull the covers off and take a look at the inner workings of SCRIBE and HyperCore it all seems complicated, however Scale has placed a very stripped down, unique GUI atop all of the complexity that is KVM. We mentioned earlier that a SMB/Mid Market’s key technology needs revolve around simplicity, availability, scale-ability and budget – let’s look further at how Scale addresses each of those factors.
You would think that utilizing KVM and and a purpose built storage solution would really complicate things for an end user – Command line and bash are not normally a familiar place in terms of an SMB customer. That’s exactly why Scale has focused a ton of effort on the GUI and overlying HyperCore software that abstracts this complexity. The interface is trimmed down, providing end users with only the functionality that they need. It’s a platform on which SMBs can deploy, manage, snapshot and protect their VMs – no selecting datastores, no configuring switches, no hardware compatibility, etc. In fact I found myself continuously asking the question “So, is this it? This is the product? Where do I get the advanced settings?” Their answer was always “Yes, this is it, there are no advanced settings, remember our target customers”. This is a hard realization for a VMware guy, someone who has been twisting knobs and playing with advanced configuration parameters for years but I do get where they are going with it. Most SMBs simply don’t have the budgets and/or time to train employees on advanced settings and just want to provision their VM and move on to the next task. With the HC3 you can accomplish just this in a matter of seconds.
Available and Scalable
Every size of business has what the market likes to call business critical applications, and this most certainly applies to SMB as well. We don’t and can’t differentiate what is and isn’t a BCA depending on business size. If we did they would be called Enterprise business critical apps. Small and Mid-Market businesses, just as Enterprise have many applications that are extremely important to their company and need to be available and protected. The HC3 provides an avenue for the SMB to gain access to what is sometimes only known as an enterprise feature; High Availability. Each and every time a VM is deployed, it’s deployed in a way that makes it highly available – no enabling anything, no tweaking knobs, this is the default behavior HC3 undertakes. Implementing disaster recovery is also a breeze using the HC3, it’s as simple as adding another HC3 cluster at a remote location, configuring it as a replication target and enabling replication on a VM by VM basis inside of the HyperCore software. DR is something that is normally out of budget to an SMB – Scale brings this within an arms reach!
As mentioned earlier SMBs are always looking to grow and when they do their technology needs can skyrocket, leaving them to answer tough questions in regards to compute, storage and licensing needs. Most often than not SMB have bought into a solution that simply cannot scale with their growth resulting in expensive rip and replace scenarios. The HC3 is built in such a way that scale-ability is on the front line, I mean, it’s right in the companies name right? When a company has outgrown their HC3 starter pack they are able to simply add nodes into their cluster, and this process is done seamlessly. Simply rack and stack another node, point to an IP inside of your configuration interface and the configuration is simply applied to the node. The HC3 can support up to an 8 node cluster, certainly providing more than enough wiggle room for most SMBs to grow.
Perhaps the biggest driver in an SMB/Mid market business is budget – they simply don’t have the dollars to spend that enterprise companies do. The HC1000 starter pack is offered with at a really affordable price, starting at roughly 25K- incredibly lower than most solutions providing the same type of functionality. There are have one licensing tier – either you have an HC3 with all of the features or you don’t. There is no paying for “enterprise” features like HA or replication. One sku gives you everything Scale has to offer. But Scale addresses budget in more ways than just money. You simply don’t need the level of employee knowledge to operate an HC3 as you would with a full blown enterprise virtualization stack, thus saving huge amounts of money and time on training. The simplicity plays a role here as well. Most if not all of the features are setup and ready to go when your HC3 arrives – no spending time on money or consultancy to configure HA or enable the migration of VMs between nodes. And the features that do require a little setup, such as replication, are completed in a couple of clicks. Support plays another key role in budget friendliness. Scale ranks way above average receiving a Net Promoter score of 79 – and they have one type – 24/7. Every HC3 starter pack that is sold comes with 1 year of support already included, 24/7, no messing around.
So in the end we are left with a company that really knows their target market and it shines through in the product they deliver. Do I think the HC3 is a good fit for an SMB – absolutely. Maybe not every SMB but I’d say 99% of them. Scale has addressed a market that a lot of companies simply skim over. Most will strip a few “enterprise” features out of their already existing enterprise product, lower it’s maximum configurations and call it their SMB offering. This is not the case with Scale – what you see is what you get – A simple, easy to use, purpose-built, scalable hyperconverged appliance that even a 4 year old can use (with the bribery of chocolate chips of course). No licensing tiers, no need for feature comparisons and tough licensing discussions, just buy what you need and grow!