Imagine having an enterprise IT infrastructure that can adapt to constantly changing needs, but without adding risk to your organization. A system that offers the predictability and familiarity of traditional IT deployments and the scalability of cloud computing models. It allocates computing resources efficiently, simplifies management and reduces the need for complex integrations. It’s one cohesive, software-defined infrastructure that can seamlessly adapt to evolving business needs, making life easier for DevOps teams and end users alike. Instead of worrying about increasing infrastructure complexity and all the security risks that go with it, businesses can finally regain control of their IT instead of having IT control them.
This is composable infrastructure. It’s a single interface that allows administrators to provision and maintain computing resources on an as-needed basis. It’s the ultimate solution for taking command of your computing assets, as well as eliminating the often poorly managed risks of cloud-hosted resources. To make that possible, IT leaders are tapping into the power of automation and artificial intelligence and doing away with repetitive manual tasks, which are vulnerable to human error. The decreased management complexity also lets IT leaders focus more attention to strategic tasks rather than getting bogged down in writing countless lines of code to integrate new applications and provision new systems.
Composable infrastructure is still a new concept, but it delivers the ultimate solution for large enterprises that need the scalability and fast provisioning of the public cloud but without losing control. It’s a solution for improving efficiency in enterprise data centers and easing management woes associated with the increasing challenges of scale, risk and compliance.
How composable infrastructure compares to other options
Most end users take access to IT services for granted, having little idea of the complexities behind the scenes. Only once something goes wrong do they suddenly realize that they can’t get on with their jobs. Morale and productivity are the first casualties of an IT infrastructure that’s unable to satisfy the needs of your employees. Those responsible for deploying and maintaining the infrastructure are often held to account when an issue arises, forcing teams to work overtime to find a resolution. At the same time, end users grow increasingly impatient and are more likely to look for inadequate and even risky workarounds. The complexities of traditional and, to a lesser extent, converged infrastructure, only exacerbate the risks.
Traditional infrastructure
Traditional infrastructure is where it all started. It’s a familiar and, in theory, highly customizable hardware-defined environment consisting of physical servers, storage and network switches. Every operating system is directly bound to hardware, and there’s no virtualization layer, which results in different systems running on different platforms with no unified management tools for crossing the divide. You end up with lots of underutilized systems leading to data silos and a serious lack of interoperability. A traditional infrastructure might be adequate for smaller environments, but once your business needs grow more demanding, it doesn’t take long before administrators and DevOps teams run into crippling complexities at scale.
Converged and hyper-converged infrastructure
Converged infrastructures started appearing around ten years ago, alongside the rapid rise of cloud computing services like IaaS (infrastructure as a service) and PaaS (platform as a service). It brings multiple physical systems, including those residing in both in-house and cloud data centers, together under a cohesive environment to offer simplified deployment and management. This approach reduces the footprint of the data center and increases efficiency by maximizing the use of available resources. Converged infrastructure may also serve as an enabling foundation for private and public cloud computing environments. Hyper-converged infrastructures take things a step further by supporting virtual workloads that don’t need to be connected to SANs, but both lack the merits of automation.
Composable infrastructure
Composable infrastructure unites compute power, storage and network infrastructure into a single platform, much like converged and hyper-converged solutions. What’s different? It uses automation and a unified API to ‘compose’ a flexible set of computing resources. It’s a modular approach to enterprise computing, but one that can be reconfigured on the fly to become whatever it needs to be. IT administrators don’t need to concern themselves with the underlying hardware, since compute power and storage are automatically requested from a pool of available resources by the API. This allows enterprises to provision workloads just as efficiently and quickly as public cloud vendors, albeit without losing control.
Choosing a composable infrastructure vendor
By delivering compute power, storage and bandwidth on the fly, composable infrastructure is a true game changer for enterprise data centers. But, since the technology is still in its infancy, there aren’t many vendors in the field. We can expect that to change over the next couple of years as automation technologies rapidly advance. The best-established vendor is Hewlett Packard Enterprise (HPE), which has been a pioneer in the field since 2015. However, they only started shipping the world’s first composable infrastructure package last year. They also released their Composable Cloud platform last November to give enterprises the ability to span composable computing workloads across hybrid cloud environments.
HPE’s biggest contender is Dell EMC with its PowerEdge MX integrated solution, which the company describes as “kinetic infrastructure.” Its unique selling point is that it allows for future hardware upgrades without disrupting operations thanks to the absence of a midplane. Other up-and-coming contenders in the space include NetApp and Juniper Networks. There are also several startups which have recently appeared on the scene, so look out for brands like Liqid, DriveScale and RackN.
How to deploy a composable infrastructure safely
The need for composable infrastructure arises from the limitations of public cloud computing. Enterprises today need the quick and easy provisioning afforded by public cloud vendors, but public clouds can’t accommodate every need. Not only are certain workloads unsuitable for the public cloud; some enterprises need to retain control over every element of their IT for information security and regulatory compliance. Composable infrastructure makes that possible using three key elements:
Establishing a root of trust (RoT)
Cyberattacks often exploit firmware which, in a connected environment, can quickly result in malware spreading across the entire network. To protect against this threat, it’s necessary to establish a root of trust that prevents servers from executing malicious firmware code. Both HPE and Dell EMC use an RoT embedded in their hardware, creating an unmodifiable fingerprint in the silicon itself. This prevents servers from booting up unless their own firmware matches the fingerprint.
Encrypting data at rest and in transit
Encryption has long been a core element in any security solution, as well as a requirement for meeting many compliance regulations. By scrambling data at rest or in transit, it will only be accessible to those with the decryption key while remaining unintelligible to any unauthorized party. Composable infrastructure can also automate data protection policies and enforce total file encryption with either public or private keys.
Physically isolating networks with air gaps
An air gap enhances security by physically isolating your data center from other networks that might be unsecure, such as the public internet. The term is analogous to the air gap used in plumbing to preserve water quality. When a network or individual device is ‘air-gapped,’ it has no interfaces with external networks, either wired or wireless. However, because air-gapping prevents any connections to the outside world, the only way to access the system is through an endpoint within the network. As such, it’s better used only for networks that handle highly classified information. Even then, air-gapped networks aren’t completely immune to cyber-risk, and there are several cybercriminal tools designed specifically to attack them.
Become more adaptable to change
Composable infrastructure is rapidly shaping up to be the next big thing in the increasingly complex world of enterprise IT. For organizations wanting to enjoy all the benefits of the public cloud without having to surrender control to a third party, it offers a promising solution that will improve businesses to be more adaptable to changing needs, and help companies regain control over their data and its security.
This article represents the personal opinion of the author.
