Introduction
The market for managed services has grown fiercely competitive, forcing Managed Service Providers (MSPs) to seek ways to remain competitive while becoming more profitable. The hype surrounding hyperconvergence—the integration of the compute, storage, and virtualization layers of infrastructure into a single solution architecture—makes the architecture seem promising. But what more do MSPs need to know to guarantee success?
Scale Computing has a decade-long track record of innovation in hyperconvergence, storage architectures, cloud-based disaster recovery, ease-of-use, and product packaging and pricing that match MSP business models. Together these innovations have given hyperconverged systems the high availability, scalability, versatility, and affordability MSPs need to become and remain competitive while increasing profits.
This white paper, intended for CEOs and VPs of Sales, Marketing, and Business Development, describes how MSPs can support more customer applications with less investment than possible with traditional data centers and other hyperconverged architectures. The content is organized into three sections:
- Highlighting some of the challenges confronting MSPs
- How the design of Scale Computing Platform addresses these challenges to give MSPs a compelling and enduring competitive advantage
- Identifying the most profitable opportunities for selling managed services and systems based on SC//Platform
Challenges Confronting MSPs
MSPs are challenged to provide reliable IT services for a reasonable price. Using many traditional technologies, MSPs face complexity, high costs, and support challenges in combining those technologies into solutions their customers need. One of the most pervasive yet challenging technologies MSPs have deployed is the traditional 3-2-1 infrastructure.
Complexity
The traditional 3-2-1 data center architecture consists of three or more servers connected by two network switches to one or more storage area network (SAN) or network-attached storage (NAS) appliance(s). This layered architecture presents some serious challenges, the most obvious being considerable complexity in bringing together the multiple layers of individual components. Using best-in-class systems is normally a best practice, but having multiple hardware and software vendors makes it necessary for staff to be trained on and use a separate management system for each layer of the stack.
Virtualization creates what should be an additional and versatile layer of abstraction, but because the stack is built with layers of components from different vendors, the complexity of the integration is not optimized for performance. These solutions usually try to overcome the inefficiencies with extra RAM and SSD caching for storage performance improvements but this consumes extra system resources that should instead be available for running more VMs.
Single Point(s) of Failure
The use of a SAN or NAS (the “1” in “3-2-1”) creates another challenge: single point of failure. Eliminating all single points of failure in a SAN/NAS solution further increases complexity—and cost. And because SAN/NAS typically has a monolithic architecture, it can be difficult and expensive to scale for capacity and/or performance. Providing disaster recovery (DR) protection can also be expensive because it often requires “doubling down” on the infrastructure footprint by deploying a fully redundant configuration at a separate site.
Some converged infrastructure solutions attempt to mimic this 3-2-1 architecture. One such approach for converging compute and storage resources relies on virtual storage appliances (VSAs) that run as virtual machines (VMs). However, because the approach is similar to the way SAN and NAS controllers function (a potentially worthy objective), VSAs usually suffer from the same resource utilization and performance problems that plague the 3-2-1 architecture (the reason why is covered in the next section).
Going Small
Beyond building their own internal infrastructure or 3-2-1 infrastructure for customers, MSPs are further challenged by delivering the right-sized infrastructure for customer sites where a multi-node cluster may be too large for their needs and budget. These may be primary sites for smaller customers, ROBO sites, or edge sites. Delivering turn-key infrastructure onsite that can be managed remotely with inherent replication back to the MSP data center for DR is more challenging with traditional infrastructure and hypervisors. These solutions often require additional levels of licensing and management tools for the MSP.
MSPs relying on traditional hypervisors and virtualization infrastructures confront these additional challenges:
- Differentiating MSP business services can be difficult when every MSP uses the same commoditized hypervisor solutions
- The enormous economies of scale in the cloud apply downward pressure on the pricing of managed services hosting and traditional on-premises infrastructure
- Requiring more highly-trained and certified professionals to manage and maintain the complexity of traditional 3-2-1 virtualization solutions, increasing costs
- Protecting data and applications from disaster and cyberattacks like ransomware can be more challenging with traditional virtualization infrastructures that are more vulnerable and harder to recover due to complexity
- Creating scalable infrastructure that can provide only the resources (and cost) desired initially, but that can grow rapidly with ease as it is needed
These challenges erode the profit potential of managed services, giving MSPs a potentially existential reason to rethink the traditional 3-2-1 architecture or poorly converged solutions based on that architecture for something more suitable for managed services offerings.
Confronting The Challenges With Hyperconvergence
With no industry standards for hyperconverged systems, all are designed differently, even using different definitions of what constitutes “hyperconvergence.” When originally coined, the word meant the convergence of compute, storage, and/or networking in a single solution that also included the hypervisor. So, the “hyper” had real meaning, not just the “hype” it is today with some solutions.
Scale Computing defines hyperconvergence as “the integration of the compute, storage and virtualization layers of infrastructure into a single solution architecture.” Scale Computing Platform architecture also adds backup and disaster recovery capabilities to the compute, storage, and virtualization layers, resulting in a fully converged, highly reliable “data center-in-a-box” solution. Such genuine hyperconvergence, implemented in a family of standalone and clustered appliances, dramatically simplifies IT infrastructure, making it easier to deploy and manage, and substantially reduces the total cost of ownership.
Scale Computing also made the design suitable for use in private, public, and hybrid cloud environments. The latter two are particularly important to MSPs because they enable offering managed services that reside exclusively in the public cloud, and those in synergistic hybrid configurations with systems deployed both on customer premises and in MSP data centers.
The remainder of this section describes some of the ways Scale Computing has advanced the state-of-the-art in hyperconvergence, with particular attention paid to those capabilities that address the challenges confronting MSPs.
Architectural efficiency forms the foundation of SC//Platform. The HyperCore™ hypervisor is lightweight with minimal overhead, making more memory available for system and application software. SC//HyperCore is a Type 1 bare metal hypervisor based on components of the ope-n source Kernel-based Virtual Machine (KVM) hypervisor. KVM has long been part of the Linux kernel and, as a result, is field-proven in both small- and large-scale deployments.
While other operating platforms typically reserve and consume 24-32GB of memory, with some consuming up to 100GB, SC//HyperCore only reserves 4GB. In a 64GB system, this results in 60GB or nearly 94% of memory being available for revenue-generating applications (vs. only 40GB or 62% for other platforms). In a 32GB system, the results are even more impressive with three times as much memory (28GB or 88%) being available for applications (vs. only 8GB or 25%).
Streamlined, high-performance storage is provided by the innovative and patented Scale Computing Reliable Independent Block Engine (SCRIBE) storage layer. SCRIBE is a carrier-class, clustered, block-level storage layer that is purpose-built to be consumed directly by the SC//HyperCore hypervisor. SCRIBE utilizes a wide-striped storage architecture that automatically discovers all storage resources, including both solid-state drives (SSDs) and spinning hard disk drives (HDDs). It then aggregates the total capacity available across the cluster and presents it to SC//HyperCore as a managed pool of shared storage. With this design, all data written to the pool is immediately available for read and/or write access by every node in the cluster.
The benefits of SCRIBE derive from the intelligent pooling of storage blocks distributed redundantly across the entire cluster, which maximizes uptime and optimizes utilization. To assure high availability by protecting against individual drive and node failures, the blocks are striped and replicated across all nodes in the cluster. And to efficiently utilize the storage pool’s total capacity, the data is automatically deduplicated.
Performance is enhanced in a variety of ways, including SSD/HDD tiering, intelligent load balancing, and elimination of the inefficiencies inherent in SAN- and NAS-based storage solutions, as well as in VSAs. HyperCore Enhanced Automated Tiering (HEAT) is a particularly powerful aspect of the SCRIBE storage layer. HEAT intelligently distributes blocks between the fast flash SSD tier and the slower, but less expensive HDD tier based on a heat map that tracks I/O operations for each virtual disk in the pool. The result is a cluster that delivers a competitive price and performance for the full spectrum of different customer needs. For applications where premium performance is warranted, the cluster’s appliances can be configured with only SSDs. Conversely, for applications that do not require the high performance of SSD, the cluster’s appliances can be configured with only HDDs.
Built-in high-availability and self-healing, complete with disaster recovery (DR) protection, and file-level data backup and recovery, combine to make any managed service carrier-class. An SC//HyperCore cluster consisting of three or more nodes provides the native redundancy and resiliency needed to ensure high availability. VMs can be replicated both locally and remotely, with the latter making it possible to provide DR protection for both the MSP cloud and systems deployed at customer premises.
The redundancy is layered and integral to the architecture and includes dual network ports, redundant power supplies, and redundant striping across the cluster’s storage pool. Built-in intelligence quickly and automatically recovers from node and drive failures by redistributing VMs and data across other nodes and drives, respectively. The system also automatically absorbs replacement nodes and drives.
Appliances can be deployed on customer premises in clusters with three or more nodes to provide local high availability, or as “single-node clusters” with drive-level redundancy to meet the needs of even the smallest sites and customers. In both configurations, VMs can be replicated to an SC//HyperCore cluster in the MSP’s cloud as part of a disaster recovery-as-a-service (DRaaS) offering or to SC//Platform Cloud Unity service (covered in the next section). This enables MSPs to quickly and easily fail over any VMs affected for protection against both site-level failures and more widespread disasters.
Once service is restored, MSPs can just as easily fail back the VMs to the customer premises. Recoveries in all cases are fast and easy. Failures in any drives or nodes are handled automatically by the system. Recovering from site-level failures requires simply cloning VMs to a secondary site. The ability to revert to a prior snapshot or mount a prior virtual disk for file-level recovery is also included. Recovery procedures require just a few clicks on the intuitive management interface, whether the problem is local to where the administrator is, at a remote customer site, or at a distant MSP facility.
Versatile scalability allows adding capacity and/or increasing performance incrementally and cost-effectively as needed. SC//Platform’s efficient architectural design allows it to run on small devices suitable for edge computing, ROBO, or small customer sites with as little as a single appliance or a cluster of small appliances for high availability and then scale out as needed. Other solutions that emulate 3-2-1 storage architectures have to start on larger systems to support their resource overhead, making them less cost-effective.
Unlike most other clustered solutions, Scale Computing permits dissimilar appliances to coexist in a single cluster, and those appliances can be configured with the CPU, memory, and storage resources needed. The more added nodes, the greater the aggregate performance and the larger the storage pool. Read/write I/O performance is further enhanced by wide striping across drives in the storage pool. By contrast, a redundant array of independent disks (RAID) requires allocating drives to a specific RAID set, which limits the ability to scale both capacity and I/O throughput.
New appliances can be added in minutes without any downtime in any existing node in the cluster. The efficient design and compact appliance form factors facilitate scaling clusters to meet the growing and changing needs within existing space, power, and cooling constraints.
Multi-cluster management of an MSP’s entire distributed, hybrid infrastructure is available (by authorized personnel) via an ordinary browser. The typical hybrid cloud consists of systems deployed both on customer premises and in the MSP’s data center(s). The intuitive design of Scale Computing’s web-based management system enables IT generalists to configure, monitor, update, failover, and failback, and otherwise operate the entire hybrid cloud infrastructure without any special training and with only a shallow learning curve. Multi-user administration supports an unlimited number of administrators, whether by shift or domain of responsibility, each of whom is assigned a unique account and login credentials.
The non-disruptive rolling system updates provide an example of the simplicity of Scale Computing’s management system. A single click is all it takes to update an entire cluster’s hypervisor, storage system, firmware, and other system software. Once initiated, the automated update process systematically live migrates VMs among all nodes to facilitate updating every node in a rolling fashion with no cluster-level downtime. Updating the entire cluster in a single, automated process has another advantage: It avoids problems caused by potential incompatibilities among different software versions.
Creating and configuring VMs is just as easy. From a single screen, administrators can spin up a VM, allocate the desired resources, and upload the operating system and other software needed to run the application. Because all SC//HyperCore nodes have access to the cluster’s entire storage pool, VM placement is determined automatically by the availability of compute (CPU and memory) resources, making it easier to “right-size” existing VMs when needed.
Scale Computing Fleet Manager is an optional cloud-based monitoring and management tool offered as-a-service to all MSPs.
SC//Fleet Manager is engineered for maximum reliability, scalability, and usability by utilizing a variety of open-source components and managed services combined with proprietary software. Unlike competing solutions, which require local installation, hosting, and ongoing maintenance, SC//Fleet Manager is provided as a highly available cloud-based multi-tenant application.
Zero-Touch Provisioning (ZTP)
The zero-touch provisioning (ZTP) feature of SC//Fleet Manager provides cloud-like ease of infrastructure configuration for MSPs, allowing them to centrally configure clusters with a few clicks and entry of the bare minimum details: cluster name, LAN and backplane IPs, and LAN gateway. All that is needed on-site is racking, cabling, and powering on, slashing configuration time and potentially removing the need for IT staff to be physically present at installation.
Nodes ordered from Scale Computing ship imaged with SC//HyperCore, which includes all necessary unique identifiers and authentication key requirements out of the box, negating the need for additional node registration or manual authentication steps required by other “zero-touch” solutions. The node simply connects, and authenticates, and if SC//Fleet Manager holds a ZTP configuration for that node’s unique identifiers, it is securely delivered and executed without any manual setup or action on the physical node, and the provisioned cluster appears in the owner’s SC//Fleet Manager cluster list ready for workload provisioning and configuration.
Secure Link
Secure Link provides MSPs with secure access to the HyperCore UI of any client cluster. When a user initiates a session, a secure outbound connection is established from the cluster to the Secure Link Server over HTTPS, which is presented in the browser to the authenticated user. All data transmitted between the browser and the cluster is encrypted using TLS 1.2.
During the Secure Link session, only the HTML of the cluster UI is transmitted to the user's browser. No lower-level data is sent or received, ensuring that only the user has access to the UI. Local SC//HyperCore authentication requirements are still in effect when using Secure Link, ensuring cluster-level authorization controls for all users. Once the browser containing the tunnel session is closed, Secure Link automatically closes the outbound connection, eliminating the access route entirely until the next use.
Secure Link eliminates the need for traditional remote access tools like VPNs or remote desktops, which can be costly and time-consuming to set up and maintain. Additionally, Secure Link ensures that there are no misconfigurations or security vulnerabilities associated with the manual setup of remote access tools, making it a more secure and convenient solution for MSPs.
It is important to note that making powerful management capabilities like these so easy to use is possible only with a hyperconverged architecture that integrates the full virtualization-server storage stack.
A Better Bottom Line
Scale Computing’s innovation goes beyond SC//Platform architecture to include how appliances are packaged and priced, which has an additional beneficial impact on MSP profitability. For MSPs, margins matter more than prices, but lower prices still matter because they lead to higher margins when using systems optimized for managed services. With SC//Platform, MSPs can “do more with less” to maximize profits through a combination of increased revenues and reduced costs.
Here is a summary of the ways Scale Computing helps MSPs grow managed service offerings more profitably:
- No licensing fee or separate service fee for the SC//HyperCore hypervisor
- No separate charges for the multi-cluster, multi-user management system
- Pay-as-you-grow pricing that becomes a “pass-through” to the customers
- Per-core pricing models that align better with MSP business models
- A choice of purchasing or leasing systems
- Ability to scale clusters and clustered resources incrementally, including by specifying the amount of CPU, memory, and storage configured in new appliances
- Speed and simplicity of scaling clusters that eliminate the need for over-provisioning
- Ease of implementation and operation that eliminates the need for training, making it possible to staff 24x7 with IT generalists
- Built-in high-availability that minimizes downtime, making it possible to offer competitive money-back service level agreements (SLAs) with confidence
- Ability to monetize high-availability and disaster recovery, creating even more opportunities for growing revenues
- Simple “out-of-the-box” deployment, remote manageability, and dependable operation combine to minimize the need for truck rolls to customer sites
Profitable Opportunities For MSPs
Managed services based on SC//Platform, complemented by standalone appliances deployed as “single-node clusters” on customer premises, can be employed in virtually all application use cases across all industry sectors. Among the more popular use cases in a hybrid cloud are:
- VMware Replacement – The high cost of VMware gives the license-free SC//HyperCore hypervisor a compelling competitive advantage
- Cloud Migrations – Tools provided by both Scale Computing and its third-party partners make it easy to migrate customer workloads to public or hybrid clouds
- Virtual Desktop Infrastructure – Its lightweight architecture and efficient resource utilization make SC//Platform systems ideal for VDI, including at smaller offices
- Edge Computing – The compact form factors, self-healing operation and remote manageability make SC//Platform systems competitive in any edge computing application
- Hardware Refresh – This routine practice affords an ongoing opportunity to sell both systems and services
- Platform Consolidation – SC//Platform’s versatility enables streamlining workloads from different hypervisors or platforms onto a single, hyperconverged solution
- Disaster Recovery – Leverage the flexibility of SC//Platform to protect workloads on multiple hypervisors and platforms
These three managed services, described in greater detail below, afford the greatest potential for peak profitability:
- Remote Monitoring and Management
- Disaster Recovery-as-a-Service
- Infrastructure-as-a-Service
Remote Monitoring and Management (RMM) appeals to customers that want to own or lease their own systems, but lack the personnel and/or expertise needed to manage them. RMM is normally a value-added service for MSPs that resell or (sub)lease the customer’s Scale Computing appliances. Each appliance is normally pre-configured and shipped to the customer site for installation without a truck roll and is then monitored and managed remotely from the MSP’s facilities. Additional value-added services include migrating the customer workloads and DRaaS.
As mentioned previously, Scale Computing’s innovative design enables SC//Platform appliances to be deployed in “single-node clusters.” The smaller form factors are ideal for smaller sites, including remote office/branch office (ROBO) facilities of larger organizations. The smaller and some of the larger appliances are also suitable for many edge computing applications.
These capabilities make SC//Platform especially competitive when selling and then remotely monitoring and managing customer premises equipment:
- Right-sized systems based on a choice of appliances and configurations make it possible to match the price to the performance needed at each site
- Quick and easy setup makes it feasible to guide an untrained user by phone through the physical installation process
- Full remote monitoring and management take over once the appliance is installed, eliminating the need to have any IT personnel on-site
- Self-healing storage keeps applications running should any drive fail, and automatically adds the replacement drive to the storage pool
- Optional DRaaS support in a hybrid cloud assures business continuity should the appliance itself experience a failure
Disaster Recovery-as-a-Service (DRaaS) can be particularly profitable because customers are willing to pay a premium for the peace of mind it affords, and providing it in the cloud is relatively inexpensive. Whether the customer owns the Scale Computing appliance or leases it as part of a managed service, built-in replication, and data protection features make DRaaS as dependable as it is profitable.
MSPs can also take advantage of SC//Platform Cloud Unity. SC//Platform Cloud Unity is implemented as a virtual appliance on the Google Cloud Platform, giving MSPs an alternative to deploying physical appliances for DR purposes, or hosting a dedicated cluster or single node. The service can be used to provide a full DRaaS offering for any SC//Platform system, whether deployed on customer premises or in MSP data centers. SC//Platform Cloud Unity provides geographic diversity in offering DRaaS offerings, being able to protect workloads to a variety of data centers. And because the service runs the same software used on the appliances, it is managed exactly as the appliances are.
SC//Platform Cloud Unity is a full-service offering that includes planning, initial and ongoing testing, data replication, continual monitoring, failover and failback, and other recovery assistance as needed during and after a disaster.
Infrastructure-as-a-Service (IaaS) is a purely cloud-based model for offering managed services from clusters deployed in MSP data centers. In this model, the Scale Computing cluster is owned or leased by the MSP, and its versatile and scalable resources are shared virtually and securely among multiple customers.
With SC//HyperCore and other aspects of SC//Platform architecture making efficient use of all compute and storage resources, IaaS often serves as a core offering that can be value-added with other services, including consulting, migration assistance and DRaaS. The combination of IaaS and DRaaS gives the customer a complete and turnkey cloud experience, while giving the MSP a predictable and potentially lucrative revenue stream. SC//Platform’s built-in high availability and self-healing features ensure that there is continuity and availability in applications, workloads, or any other IaaS offerings that customers depend on.
Getting Started
The efficient architecture, the highly reliable high-performance storage pooling, the self-healing high availability, the seamless and incremental scale-out, the non-disruptive updates, and the ease of implementation and operation all combine to give SC//Platform its industry-leading profit potential for providing managed services.
Scale Computing appliances have been used in both the smallest and the largest deployments of hyperconverged configurations in the world. The smallest is ideal for “single-node clusters” at customer sites, while the largest gives MSPs the confidence needed to grow services without limits. The extensible architecture, which facilitates adding new and enhanced capabilities, combined with Scale Computing’s proven track record of continuous innovation, provides MSPs the assurance of a future-proof solution. And when needed, Scale Computing backs the entire product line with responsive, world-class support.
To help you get started and grow profitably, Scale Computing offers a variety of enablement and engagement resources. Enablement resources include sales and technical training materials, and regional technical, sales, and channel personnel. Engagement resources include a Market Development Fund and other marketing programs, customizable marketing and sales materials, sales leads, sizing and configuration tools, and hundreds of customer success stories, at least one is certain to resonate with any prospect you encounter.
To learn more about our MSP Program and how your business can grow profitably and otherwise benefit from SC//Platform, or email channel@scalecomputing.com, or call 877-722-5359.