Scale Computing HC3
What sets Scale Computing HC3 apart from other HCI solutions available on the market?
Scale Computing HC3 is a hyperconverged infrastructure(HCI) solution that has been hitting the IT market lately. In 2019, Scale Computing received top CRN ARC awards in the Converged/Hyperconverged Infrastructure and server and desktop virtualization categories, distancing many very well-known competitors in the market.
What is a true HCI solution?
You've probably heard of hyperconvergence or hyperconverged infrastructure (HCI). These terms are becoming increasingly popular in the IT world. However, what do they really mean? Do vendors who claim to have hyperconverged infrastructure really meet the conditions of hyperconvergence?
Hypervisor
The term hyperconvergence refers to the incorporation of hyperconverged (supervisor, manager) virtualization into a complete solution that includes storage (disk), server and virtualization. Some people think hyperconvergence is a synonym for "super convergence," but that's not what's meant here. "Hyper" in convergence means hypervisor.
Many solutions that are called hyperconverged use third-party hypervisors such VMware or Hyper-V for virtualization. However, this does not fit the definition of true hyperconvergence.... A true hyperconverged solution with a native and fully integrated hypervisor allows for tighter integration of components such as hypervisor-embedded storage architecture, better automation of management tasks and self-repair capabilities.
One of the benefits of true HCI, very important for users, is independence from virtualization system vendors and their imposed licensing methods. True HCI is a complete virtualization and infrastructure solution from a single vendor.
Is Scale Computing HC3 a true HCI?
Scale Computing HC3 is a true HCI solution with its own proprietary operating system and HyperCore hypervisor, tightly integrated into the storage architecture under the SCRIBE name. The Scale Computing HC3 solution comes as a complete appliance that integrates all the components required to run virtual machines, with the ability to expand if requirements increase.
Scale Computing HC3 not only provides a full virtualization infrastructure, but also includes the HC3 Data Protection Suite with features such as snapshot copies, remote replication, switching to a remote location and vice versa, and options for recovering individual virtual machines (VMs) or files.
The real benefits are simplicity, storage efficiency, application flexibility, from data centers to small remote branch offices.
Scale Computing HC3 is simpler?
Simpler does not always mean better. IT pros would always like to have access to buttons, knobs for every possible option, and sometimes they are indeed useful for certain tasks. But when it comes to everyday tasks, fewer buttons to press and fewer consoles to monitor can save valuable time that can be used for other tasks.
Scale Computing HC3 simplifies IT management through the use of automation and simple workflows, starting from deployment, throughout the life of the solution.
HC3 implementation vs. other HCI solutions
Scale Computing HC3 is designed as a complete virtualization solution that allows users to move as quickly as possible from unpacking hardware upon delivery to creating virtual machines. Skipping the step of mounting the hardware in a rack and cabling it, configuring the appliance only requires specifying IP addresses, entering a license code, and then redirecting it in a web browser. Creating a cluster requires only indicating the IP address of the first node on additional nodes using the "join" command.
Unlike other HCI solutions, no manual configuration of disk storage is required. No hypervisor or separate management console installation is required. A three-node HC3 cluster can be rack-mounted, wired and configured in less than an hour; including running the first virtual machines.
Management and monitoring
Scale Computing HC3 offers a single management interface for managing servers, virtualization, storage and data protection. HCI vendors using third-party hypervisors require a single console to manage virtualization, and need separate consoles to manage storage.
Condition monitoring machines and self-repair
HC3 's unique architecture includes active state machines that monitor each device and are able to take an automatic repair action or at least report the occurrence of an error or failure. The automated, self-repair actions taken by the state-monitoring machines and the high availability provided by hypervisor clustering enable system continuity in the event of an error or failure.
Scale up and scale out Scale Computing HC3
Expanding a Scale Computing HC3 cluster is as easy as its initial deployment, even when the newly added node to the cluster has a different configuration of CPU, RAM and disk storage. Many other HCI solutions require that new additions to the cluster have the same or very similar configuration as existing nodes.
Snapshot copies and cloning
Snapshot copies of HC3 VMs are used in HyperCore hypervisor for data protection, replication AND fast VM restoration. Snapshot copies (snapshots) in Scale Computing HC3 are used very often, and because they are made very efficiently it is possible to store up to 5000 snapshots for a single VM.
Unlike other virtualization solutions, VM clones in HC3 are not bound by a "parent-child" relationship with the original VM. Clones are created in a "thin" manner sharing data blocks with the original VM. Clones, however, are completely independent, and the removal of the original VM or other clones has no effect on the cloned VM.
What distinguishes Scale Computing HC3 disk storage ?
Due to the need for tight integration of hypervisor, hardware and storage, Scale Computing has created a hypervisor-embedded storage architecture characterized by high performance, yet high simplicity. This unique architecture provides numerous benefits.
Automatic configuration
Other HCI solution providers require manual creation of disk memory groups from existing memory before creating virtual machines. Some, software-only HCI solutions require manual memory configuration before installing the HCI solution itself, followed by manual configuration related to hypervisor requirements. Manual management activities are primarily related to the need to support third-party hypervisors.
With Scale Computing HC3, the storage is embedded in the hypervisor and is automatically configured across the cluster. No manual configuration is required for deployment and subsequent management. All disk storage in the cluster is combined into a single storage pool, used by all nodes in the cluster. As new nodes are added to the cluster, their disk memory is transparently and automatically added to the disk memory pool, with no manual configuration required.
SCRIBE
The Scale Computing Reliable Independent Block Engineis the core component of the Scale Computing HC3 system, combining disk drives from each HC3 node into a single, logical pool of disk storage. Pooling occurs automatically without any configuration. Blocks are written redundantly across the system, allowing for the loss of individual disks or even an entire system node.
The SCRIBE disk storage pool is available to all system nodes and is provided without any file systems, protocols or virtual appliances. SCRIBE is built directly into the HC3 HyperCore operating system. When VMs are created on HC3, virtual disks provide direct block access for VMs to the SCRIBE disk storage pool. The only file systems created on Scale Computing HC3 are the file systems used by guest operating systems for VM access to virtual disks.
Other disk storage architectures try to emulate SAN or NAS devices, traditionally used in virtualization. They start with a disk storage pool at the lowest level with a file system installed, then shared with a hypervisor, which creates yet another file system managed in effect by a virtual storage appliance (VSA), to finally share disk space with VMs that create yet another file system. The VSA that manages disk storage, in addition to the many levels of protocols that each I/O operation must go through, consumes a large portion of RAM (often in the range of 24GB+ per cluster node) that could be used to create more VMs.
For comparison, the route of I/O (input/output) operations in the VSA architecture can look as follows:
HEAT
The main feature of data sets is that only a small percentage of them are actively used data. In a hybrid, tiered disk storage system, the greatest performance benefits are achieved when active data is stored on SSDs and inactive data is stored on slower, traditional HDDs. HEAT technology monitors data access and creates a dynamic map of active data blocks, then moves those blocks to the SSD memory layer, while moving inactive memory blocks from SSD to HDD.
Moving data blocks between layers is done automatically, transparently to users and even administrators.
For each virtual disk, the relative priority of SSD usage can be adjusted from 0-11. The default priority for all disks is 4, and if no disk is ever changed, no disk will have priority over another. The priority of any disk can be dynamically increased or decreased, increasing or decreasing SSD memory usage for that disk. Priority 0 means skipping SSD memory altogether and using only HDD storage. In the case of priority 11, the system will try to put all the data of this disk on the SSD memory, or at least in the available SSD memory taking into account the priorities of the other disks.
Scale Computing HC3 has built-in, real-time IOPS (input/output operations per second) meters for each virtual disk. The results of changing the priority for a specific disk can be observed immediately. Because virtual disks can change dynamically, priorities can be adjusted to specific needs to achieve the right balance of IOPS performance between virtual disks and virtual machines.
What is the difference between HC3 and Edge Computing?
Edge computing describes physical computing infrastructure that is intentionally placed outside the "four walls" of the data center. The goal of edge computing is to place applications, computing resources and storage close to where they are needed, used and where data is collected. Outside the data center, the infrastructure requirements for edge computing can be much smaller, if not negligible, when comparing it to a typical small server configuration.
Adequate infrastructure
Storage architectures using VSA devices offered by other HCI solutions can consume large amounts of RAM and even CPU. These solutions cannot scale down to smaller devices/server due to high resource consumption.
Compared to other hyperconverged virtualization solutions, HC3 HyperCore is extremely "lightweight," mainly because of its storage architecture. HyperCore, with its built-in SCRIBE disk storage layer, allocates only 4GB of RAM to run. This has allowed Scale Computing to offer the HC3 edge processing HE150 appliance with just 8GB of RAM.
Other HCI solutions that require a virtual storage appliance (VSA) that are actually VMs consume more than 24GB of RAM right away before VM creation even begins.
Total cost of ownership
For edge computing deployments, there may be hundreds or thousands of sites requiring hardware infrastructure.
The cost of such a deployment, due to unnecessary, excessive resource consumption and overly large deployment scale, will be enormous, unacceptable especially if fault tolerance is required. In the case of Scale Computing HC3 and appliances such as HE150, a high-availability, three-node cluster can be deployed at the same cost as a small server or HCI appliance from another vendor.
Stories of Scale Computing HC3 deployments successfully completed
There are a very large number of successful deployment stories available, describing use cases in almost every business domain around the world. They describe the problems customers faced, why they trusted Scale Computing HC3, and how HC3 solved their problems.
A full list of these documents is available on the Scale Computing's website.