Enhancing Your Composable Infrastructure with Lightbits

  • What is Composable Infrastructure
  • The Benefits of Composable Infrastructure
  • Composable Architecture: Disaggregated and Scalable Storage, a Key Building Block
  • Composable Infrastructure and the Cloud
  • Lightbits Composable NVMe/TCP Storage

What is Composable Infrastructure

Traditionally applications run on servers that have been pre-configured with specific compute, networking, and storage resources. Composable infrastructure is a modern data center architecture that decouples applications and workloads from the underlying physical resources. This type of architecture pools your physical resources and makes them accessible over the network fabric, so you can easily and efficiently provision them on-demand to your applications.

The Benefits of Composable Infrastructure

A composable infrastructure gives you the ability to dynamically configure and manage via software the environment used to run your workloads. You can compose on-the-fly a different environment for each workload by taking compute, network, and storage resources from pre-defined resource pools. In this way you get full flexibility to compose a different and optimal environment for each workload based on the workload’s specific needs, avoiding under-utilization and over-provisioning. When you complete running your workload, the resources are returned to the pools and can be used later to compose new environments for other workloads. Sounds easy? Well, it depends if you choose the right building blocks.

Composable Architecture: Disaggregated and Scalable Storage, a Key Building Block

Regardless if you run on a public cloud or on-premise, no matter if your workloads run as virtual machines, as containers, or as traditional bare-metal applications any composable infrastructure requires a pool of storage and the ability to dynamically provision it across different workloads. No need to deploy dedicated, costly, and complex storage networks. No need to get into the scale-hell of hyper-converged infrastructures. All you need is to disaggregate your storage and share it over your existing IP network already accessible by your workloads. You can move your workloads anywhere they have IP connectivity, letting them run on any compute pool totally unconstrained from the storage. Easy, efficient, flexible.

While the composable infrastructure can provision your disaggregated storage on-demand, your storage needs can grow over time. The ability to scale out your storage pool is a must-have capability to make sure your composable infrastructure can grow as well. Last but not least, as you grow and scale, failures will happen more frequently. Efficiently protecting your data from (un)expected failures at scale becomes critical to let your composable infrastructure run smoothly.

Composable Infrastructure and the Cloud

Composable infrastructure can be used as the foundation for any private, public, or hybrid cloud. It pools your physical resources and makes them available on-demand to run the applications. A composable disaggregated architecture allows IT data center operators to achieve the flexibility, cost, and benefits of cloud but on-premise.

Lightbits Composable NVMe/TCP Storage

Lightbits composable storage lets you create a pool of disaggregated NVMe storage using the servers you choose. You can dynamically provision storage over standard TCP/IP networks, so your workloads will have storage access from any place they may run or move to.  The storage servers work in concert to create a cluster and protect your data. In case of failures, Lightbits software will rebalance and re-protect your data across the cluster in the background, transparently to your workloads and transparently to your composable infrastructure management software.

Need more storage for your composable infrastructure? Just add more servers to the existing Lightbits cluster or more SSDs to the existing storage servers. You choose your preferred hardware vendors.

You may be asking why you need the high IOPs and the low latency of  NVMe and NVMe/TCP for your composable infrastructure. If you ask, it’s because you don’t run I/O intensive applications, otherwise, performance is the simple answer.  But if you don’t run such applications, performance is still the answer. At scale, when you run hundreds or thousands of workloads in your composable infrastructure,  the aggregated performance can become really high, even if each individual workload needs a low portion of the overall requirements. The higher the performance of your storage, the more you will be able to scale your compute pools without being forced to acquire more resources to improve your storage performance.

Additional Resources:

About the Writer:

Chief System Architect