Too often, storage is relegated to secondary status in network planning. But with data volumes exploding and edge computing expanding, it’s time to reevaluate network storage and deployment.
Here is a list of best practices for tuning up storage on the network.
1. Work Closely with the Storage Team
Network teams should have strong collaboration with the IT personnel responsible for storage purchasing, allocation and architecting. Storage decisions in IT are most often made around how much storage the organization needs, the types of data the site needs to store, how quickly teams need to access data and storage security.
A storage manager likely won’t look at network configurations around storage, or how differences in network speeds and priorities can affect storage performance. Unless teams take a holistic look at storage access, security and data flows throughout the network, storage, data and applications lenses, storage needs are likely to be misconstrued. Will storage managers adopt this holistic view? History tells us they tend to act somewhat independently. It’s up to network managers to develop a useful collaboration.
2. Understand the Basics of Storage Architecture
As organizations deploy more edge networks, they distribute storage to multiple sites throughout the enterprise, while maintaining their primary high-capacity data storage infrastructure within traditional internal data centers and cloud environments.
Regardless of a company’s storage deployments, storage managers generally work from a basic three-tier architecture. Network managers should be aware of this architecture when discussing network storage with storage managers.
The basic storage architecture consists of the following three-tier structure:
-
Solid-state drives (SSDs) are used for rapidly processed transactions that require optimal throughput.
-
Fast hard disk drives (HDDs) support systems where transaction throughput is less demanding than it is in SSD environments, and where certain batch data processes are performed, usually executed intraday.
-
Slow hard drives and tape drives are used to store and retain seldom-accessed data that doesn’t require rapid access speeds. These drives are also used for system backups and recovery. They are cheap to buy and install.
Storage architecture also tends to use system-on-a-chip (SOC), which is typically found in smartphones and other IoT devices. The SOC chips enable a modicum of data storage as well as a small CPU and GPU to reside on a single chip in a device. These chips work in concert with other mobile storage technologies, such as low-power double data rate (LPDDR) and synchronous dynamic RAM (SDRAM).
4. Hard Disk Drives are Still King
Despite the advent and falling prices of solid-state storage and SOCs to improve transaction throughput and the capabilities of mobile devices, hard drives are still king in data centers, cloud environments and server storage. Hard drives have a long history of reliability, and the technologies behind them continue to move forward. They are a dependable way to store terabytes of data anywhere — and for a reasonable cost.
Network-attached storage (NAS) is a prime example. Although NAS can be deployed with solid-state storage, teams most often deploy it with hard drive storage that can be configured by itself as a single disk, or in a redundant array of independent disks (RAID) array that improves performance and creates backup capabilities.
5. Discuss Auto-Tiering with Storage Managers
The network has an assortment of automation tools that can route traffic based on network throughput circumstances. These tools, based on rules set by network staff, can automatically allocate or deallocate bandwidth, effect failover and so on.
A separate set of automation rules and tools is likewise used for computer storage, called auto-tiering. Auto-tiering moves volumes of data between the various SSD and HDD storage tiers in ways that minimize costs and deliver consistent performance and data throughput.
Because network automation does much the same in managing network traffic and throughput, it’s in the best interests of network and storage managers to collaborate on their automation rules so network automation doesn’t conflict with storage automation.
Working together, network and storage managers can create best-performance and least-cost strategies that work for both teams. Few sites have network and storage managers working together on automation goals and rules, but they should. What network managers can do now is start the process. Arrange a meeting with your storage manager, with the goal of synchronizing network and storage automation for best results.
Speak Your Mind