The more the enterprise gravitates toward software-defined architectures, the more it is confronted with diverse data loads and increasingly complex application requirements. This is leading to a Catch-22 in that while overall hardware requirements are diminishing, the enterprise still needs to field a wide variety of solutions in order to provide optimal support for emerging workloads.https://o1.qnsr.com/log/p.gif?;n=203;c=204663295;s=11915;x=7936;f=201904081034270;u=j;z=TIMESTAMP;a=20410779;e=iOne of these challenges is increasing the availability of in-memory and on-server storage, which goes a long way toward removing latency in high-speed applications. According to a new study by storage subsystem provider Crucial, two thirds of IT decision-makers say they need to expand their capacity of server-side memory in order to support the increasing number of virtual machines under management. As in the traditional server farm, the need for more server memory is driven by the abrupt spikes in data that virtual machines tend to generate, not that the average load is pushing the limits. In the survey, 58 percent of respondents said they typically use less than 60 percent of the memory they have now.
On-server solutions also present another problem: They only provide lightning-fast service to the server they are attached to. This is problematic for many virtualized workloads because it diminishes the ability to consolidate resources or pull data from machines hosted elsewhere in the data center. Companies like Plexistor are targeting this problem, however, with advanced fabric technologies designed to maintain memory speed across distributed architectures. At a recent demo, the company showed its Persistent Memory over Fabric (PMoF) Brick architecture and 100 GbE connectivity hitting 1.6 million random 4k IOPS at less than 6 microseconds, while maintaining throughput of 7 GBps. In this way, the company says it can dramatically out-perform leading rack-based SSD solutions, which typically offer latency of about 100 microseconds.
But don’t write Flash off quite yet. A company called Datrium recently released a new system that enables pooled storage that scales up to 100 TB per server with performance hitting upwards of 3 million IOPS. The company’s DVX software is designed to integrate into legacy data center infrastructure, providing high-performance, high-density architectures that rival more costly and disruptive converged and hyperconverged solutions. The system also allows for data acceleration on demand, plus the ability to provision speed separately from capacity by preserving IO on the local server while maintaining shared durable storage in an off-host repository.
Meanwhile, a company called Cohesity is out with a new policy-based automation platform designed to increase availability of both primary and secondary storage in VMware environments. The system combines the Cohesity DataPlatform and DataProtect solutions with VMware’s Virtual SAN and vRealize Automation to enable converged solutions that maintain instant recovery, automated data protection and other features without the need for expensive and complicated dedupe and backup infrastructure. With this level of coordination between server-side and distributed storage systems, the company says it can enable a high degree of flexibility within private cloud architectures by automating replication and archival policies within the VM provisioning process.
Improving storage performance, then, is more than finding a better way to save and retrieve data. If handled correctly, it can boost performance across server and networking resources as well, leading to a more efficient and effective data environment.
And ultimately, this will help the enterprise tailor infrastructure and architectures to an ever-increasing variety of data needs.
Arthur Cole writes about infrastructure for IT Business Edge. Cole has been covering the high-tech media and computing industries for more than 20 years, having served as editor of TV Technology, Video Technology News, Internet News and Multimedia Weekly. His contributions have appeared in Communications Today and Enterprise Networking Planet and as web content for numerous high-tech clients like TwinStrata and Carpathia. Follow Art on Twitter @acole602.