Even with many cumulative years of storage technology advancements, infrastructure management has only become complex, especially for IT organizations that must grapple with inflexible operational budgets.
Across all industries today, many companies are turning to automation to address these challenges, exploiting machine learning and data-driven artificial intelligence methods. Indeed, the most urgent needs for the storage industry are simplicity and automation. Specifically, this aim is to simplify and automate data management across a range of heterogeneous storage configurations—and also improve the ROI for operational storage infrastructures. These needs are being met with advances in software—especially in microservices and cloud-native architectures.
In this article, we briefly examine the evolution in computing, networking, and storage sectors, and also predict what is likely to come in the future.
The compute revolution — virtualization has led to containerization and serverless technologies
In ancient times, all we had were physical servers, which would be racked and stacked. Carefully and tediously, it was necessary to layer on a cumbersome operating system and specific software to perform specific tasks on a specific server. Then, virtual machine technology became a major industry force. While a VM resides on a physical server, it runs on a hypervisor operating system that can support multiple VM or virtualized servers. This architecture makes it possible to run various workloads simultaneously—on a single physical server. This enables far better capacity utilization, and offers a high-speed option for instantiating new workloads. Mass migration to the cloud actually began with virtualization, and it has made moving to the cloud more easily achievable for many companies.
Organizations that stubbornly maintain their IT infrastructure entirely on-premise must continue to wrestle with the challenges of resource, cost, and management restrictions. For those with enterprise-level aspirations, their ability to scale will be severely limited. Some organizations will continue to maintain large, on-premise data centers, out of fear in venturing into the unknown. Many other businesses—hampered by the limitations of virtualization—are in the process of migrating their systems infrastructure from their physical data centers to the cloud. So, where does this leave the virtualization movement?
Containers in the Cloud
Containerization is the next advancement beyond virtualization. Containers provide a compute environment that is much less cumbersome. Typically, a container is much faster on startup than a virtual machine, doesn’t require an extensive OS, requires less maintenance, and is readily portable from on-premise to cloud platforms. A major difference between virtual machines and containers is that the container doesn’t have nearly the same amount of operating system overhead. The container manager uses many native OS resources. And, the container manager can run either in a physical box or virtual machine.
Containerization movements such as Docker continue to generate plenty of interest. Though there is plenty of deployment waste and inefficiency, actual container usage continues to rise steadily. The goal of many IT organizations that invest in containerization is to address the various challenges of cloud computing. Where possible, some companies are making progress with container innovations.
Networking — Breaking free of conventional hardware constraints
The best networks are very fast—and have the highest availability and reliability. These networks are built with spec hardware and custom silicon (ASICs). The larger the box, the more a vendor will charge, which encourages vendors to build network hardware even larger, highly complex, monolithic systems. As any seasoned technology professional knows, such systems lack flexibility, which has contributed to the rise of SDN and NFV.
SDN and NFV — The most recent evolution in networking today
Advances in spec hardware and white-box networking—together with significant improvements to standardization and developer tool—a major technology shift is underway in networking software. This undercurrent is now supporting all viable SDN and NFV technologies. Today, network software is now cleanly decoupled from the hardware, so that it’s no longer constrained by the box that is hosting it. Portable, reliable network software is the reason why SDN and NFV are now essential to the best available network architectures.
SDN and NFV benefits boost innovation, since organizations can create entirely new types of business models, applications, and services. Also, companies can offer new revenue-generating services. With SDN, you can operate with agility to quickly deploy new applications, infrastructure, and services—in response to ever-changing requirements.
Another rising trend in SDN is the software-defined wide area network (SD-WAN). This is an extension that applies SDN to network connections over a wide geographic area. Enterprises are increasingly integrating SD-WAN because it provides a number of benefits. These include unifying most enterprise network connections, more robust security through end-to-end encryption, and cost-reductions by eliminating the need to maintain so many expensive routers. This trend recursively affects SDN and NFV, since it reshapes the modern technology perspective and reinforces the growing demand for SDN and NFV.
Two other aspects of the infrastructure revolution include VSAN and cloud-native storage. Let’s briefly take a look at each.
VSAN has gone a long way
There are two primary goals with Virtual SAN (VSAN). The aim is both to lower storage costs and also to eliminate the I/O latencies that are inherent in networked storage. VSAN achieves fairly high availability by replicating storage objects across the cluster. These objects include snapshot images, VM disks, and VM swap disks. Storage object replication enables administrators to specify per-VM failure tolerance— with nodes, drives, or the network. Also, it addresses latency by employing flash-memory storage devices for read-caching and write buffering. There is also support for 10GbE connectivity. Another benefit of VSAN done well is its relatively low-barrier storage and lower dependence on storage stacks. Moreover, there is some progress towards hyper-convergence—in which virtualization fully integrates with compute, storage, and networking.
The limitations of VSAN are well-known. Scalability is a serious challenge. Many data centers require the intuitive manageability and proper scalability that a virtualized infrastructure offers. Therefore, these data centers need storage solutions with full capability to seamlessly integrate into virtualized environments. Too often—on the way to lowering storage costs with VSAN—an unintentional consequence is that server-side storage actually increases both capital expenditures and operating costs. Whenever there is a need for additional storage capacity, server-side storage architectures are hobbled by the necessity of physically attaching storage to the hypervisor itself. This is a highly frequent need, as storage growth rates continue to rocket upward. It may be sensible for infrastructure planners to pre-install flash, SSD and HDD capacity into their virtualized server infrastructure. However, this typically entails a huge initial investment. And, since the storage costs continue to drop with the passage of time, IT planners have strong incentive to wait—deploying resources only at the moment they are requested.
Software Defined Storage — Moving away from flash-SAN arrays
Software-defined storage (SDS) is another rising trend, though its concepts and value remain unclear for many IT directors. SDS benefits organizations of all sizes, in many ways. Moving into the future, SDS should prove to be a storage technology of choice for data centers around the world.
Software-defined storage delineates the control layer from the data layers. This enables administrative functions that are normally tightly bound to storage hardware to be available through 3rd-party software. This decoupling adds significantly flexibility and simplicity, and offers other key benefits. A primary advantage of a solid SDS solution is its versatility. SDS can adapt to various use cases and different environments. Also, performance and reliability tests are proving that properly configured SDS can easily exceed flash-array performance levels.
Advantages of cloud-native storage for containerized apps
Cloud-native architecture will perhaps be the foremost juggernaut trend in the infrastructure revolution. Such architectures are horizontally scalable across many servers. There is no single point of failure, it achieves the highest levels of resiliency and self-healing, and there is minimal operational overhead. Little, if any, human administration or oversight is necessary, since all interaction is API-driven and quite insular from the supporting platform and hardware.
The time is now upon us, when all of the benefits of cloud-native architecture are carrying through to cloud-native storage. Today, the storage revolution is still underway. Container technology solves many problems for managing cloud applications. There are a number of various approaches to deployment across the storage landscape today. These include proprietary storage arrays, centralized file systems, distributed storage, public cloud SANs and API frameworks.
The best available solutions today would involve a software-defined, platform-agnostic, orchestrated storage system that scales horizontally and provides Kubernetes or Docker integration. This would enable organizations to run enterprise-grade, containerized production apps that are consistently high-performance and highly available.
Into the future with a full-featured, cloud-native storage platform
Reduxio architecture enables unmatched scalability and deployment flexibility. We achieve this with for major design elements:
- Cloud-native storage platform — our SDS platform is specifically built for enterprises and cloud service providers. It delivers public-cloud scalability at an excellent value.
- Unified data management — Primary and secondary data management is readily accessible in a single platform. Benefits include built-in CDP, copy data management, disaster recovery, data migration, and cloud data synchronization.
- Microservice architecture — Three microservice subsystems enable independent scaling of storage capacity and performance and capacity—with unparalleled efficiency.
- Enterprise-grade performance — Combining our cloud-native microservices architecture with the very latest advances in hardware and software results in extremely low latency and the highest throughput.
Cloud-native solutions go well beyond to provide the highest degrees of elasticity, efficiency, and flexibility—at the lowest acquisition and operational costs.
The cloud-native platform from Reduxio provides robust support for edge, private, public, hybrid, and multi-cloud deployments. With cloud-native storage, you can achieve serious reductions in global data volume, advanced data mobility, active tiering, integrated data protection, and instant recovery.
Your organization can readily apply all the cutting-edge benefits of microservices to your storage infrastructure.
Each platform component is independently deployable and upgradable, and new capabilities can be integrated without the need for a forklift upgrade. Reduxio gives you the power to independent scale specific applications, enhance system performance, and maximize capacity to ensure optimal efficiency for your infrastructure.
Want to learn more? Contact us with any questions you may have about how best to leverage our leading cloud-native SDS storage and data management platforms.