Part flash, all flash or dedicated flash?

Location, location, location


Flash memory is rarely out of the news these days as recent announcements from Texas Memory Systems and EMC show.

We're now seeing a proliferation of flash across the enterprise, from PCI-E solid state drives (SSDs) in servers to solid state arrays. The options seem overwhelming and the boundaries between the different storage array options are becoming blurred. Here's a quick overview of what is in the market today and how they all fit together.

Part flash

There are two ways to implement a part-flash array today. The first is to use solid-state drives as part of a tiering strategy. SSDs were slow to be adopted into a tiering model because they were expensive and it wasn't easy to determine for a particular host which LUNs would be the hot ones and which would be relatively inactive. The initial solution was to partition up the host or assign it all flash LUNs, techniques that were both time-consuming and expensive.

Block level tiering changed that; this feature – now available from vendors like IBM, EMC, Hitachi, HP/3PAR and Dell/Compellent – breaks down a LUN into blocks, placing each block on a tier of storage based on activity. In this way, only "hot" active blocks of data sit on SSD.

This is a more efficient solution than standard tiering but is not well optimised to variable workloads where the hot blocks change frequently; the lag time of the array rebalancing to improve performance may negate any benefit (Note: without going into technical detail, solutions such as Compellent can circumvent some of these issues).

The second option is to use flash as a cache for writes. This is the solution used by arrays that make use of the ZFS file system from Oracle. Writes are written to mirrored SSDs using a logging mechanism called ZIL (ZFS Intent Log). They are then de-staged to main spinning disk asynchronously.

Reads are cached in memory using an SSD cache called L2ARC (level 2 Adaptive Replacement Cache). Write performance on the ZFS solution is directly proportional to the amount of cache available. This results in a tradeoff between the cost of cache and the level of write activity on the array.

This leads nicely into discussing solutions that potentially offer better performance but at a greater cost.

All flash

Today's solid-state drives have been designed to emulate and to be plug compatible with standard hard drives. It is therefore possible to create an array based purely on SSDs alone. This idea would be extremely expensive (EMC has announced\ it will provide these devices to customers), but do they really guarantee better performance? In the last 20 years since the first Symmetrix came to market, all storage array vendors have worked tirelessly to improve I/O performance in order to mitigate the shortcomings of the main component – hard drives.

HDDs I/O has both latency and seek time delays, which can vary depending on the workload type. Techniques such as striping, mirroring and caching have all been developed and honed to squeeze the best out of a relatively slow storage medium. If we now add solid state into the mix, response times from these devices will be many times faster than those of traditional drives and it is likely that this performance will exceed the array's ability to cope with the I/O load.

Where HDD response time was once the bottleneck, we could see I/O paths, cache and processor performance being the limiting factor. The cost/benefit calculation becomes more complicated to perform, as the limit of performance isn't as easy to predict. Of course if any vendors out there want to send me a solid state array to test, feel free and I'll answer the question for you. To my knowledge EMC is the only vendor certifying and offering full-flash arrays for their traditional VNX & VMAX ranges.

Dedicated Flash

Not all flash arrays on the market are the same. We have seen a new breed of products from Violin Memory, Texas Memory Systems and Kaminario that use flash and DRAM to create storage arrays specially tuned to work with SSD. This means being able to cope with the throughput solid state drives can offer but also being aware of the difference in technology between solid state and traditional hard disks. For example, solid-state drives (and especially MLC drives) have a limited lifetime.

Techniques such as wear levelling are used to mitigate against this and improve what is known as write endurance. Violin Memory takes this a step further and wear level across the entire array, ensuring SSDs last as long as possible. There's no doubt that dedicated flash arrays will have the fastest performance. They have been purpose-built to make use of SSD and have a price point that is only justified by the extreme performance they offer.

Summary

So what's the best approach? Well as usual "it depends". Understanding workload profiles and demands is a must; determining whether I/O latency is a restricting factor on application performance is tricky; but most important is determining whether reduced I/O times will result in significant improvements at the application layer. The chances are an SSD solution will provide improvements. The question is: is the increased cost justified? ®

Chris M Evans is a founding director of Langton Blue Ltd. He has over 22 years' experience in IT, mostly as an independent consultant to large organisations. Chris's blogged musings on storage and virtualisation can be found at www.thestoragearchitect.com.

Similar topics

Broader topics


Other stories you might like

  • Why Wi-Fi 6 and 6E will connect factories of the future
    Tech body pushes reliability, cost savings of next-gen wireless comms for IIoT – not a typo

    Wi-Fi 6 and 6E are being promoted as technologies for enabling industrial automation and the Industrial Internet of Things (IIoT) thanks to features that provide more reliable communications and reduced costs compared with wired network alternatives, at least according to the Wireless Broadband Alliance (WBA).

    The WBA’s Wi-Fi 6/6E for IIoT working group, led by Cisco, Deutsche Telekom, and Intel, has pulled together ideas on the future of networked devices in factories and written it all up in a “Wi-Fi 6/6E for Industrial IoT: Enabling Wi-Fi Determinism in an IoT World” manifesto.

    The detailed whitepaper makes the case that wireless communications has become the preferred way to network sensors as part of IIoT deployments because it's faster and cheaper than fiber or copper infrastructure. The alliance is a collection of technology companies and service providers that work together on developing standards, coming up with certifications and guidelines, advocating for stuff that they want, and so on.

    Continue reading
  • How can we make the VC world less pale and male, Congress wonders
    'Combating tech bro culture' on the agenda this week for US House committee

    A US congressional hearing on "combating tech bro culture" in the venture capital world is will take place this week, with some of the biggest names in startup funding under the spotlight.

    The House Financial Services Committee's Task Force on Financial Technology is scheduled to meet on Thursday. FSC majority staff said in a memo [PDF] the hearing will focus on how VCs have failed to invest in, say, fintech companies founded by women and people of color. 

    We're told Sallie Krawcheck, CEO and cofounder of Ellevest; Marceau Michel, founder of Black Founders Matter; Abbey Wemimo, cofounder and co-CEO of Esusu; and Maryam Haque, executive director of Venture Forward have at least been invited to speak at the meeting.

    Continue reading
  • DataStax launches streaming data platform with backward support for JMS
    Or move to Apache Pulsar for efficiency gains, says NoSQL vendor

    DataStax, the database company built around open-source wide-column Apache Cassandra, has launched a streaming platform as a service with backwards compatibility for messaging standards JMS, MQ, and Kafka.

    The fully managed messaging and event streaming service, based on open-source Apache Pulsar, is a streaming technology built for the requirements of high-scale, real-time applications.

    But DataStax wanted to help customers get data from their existing messaging platforms, as well as those who migrate to Pulsar, said Chris Latimer, vice president of product management.

    Continue reading
  • Infor to stop developing on-prem software for IBM iSeries
    ERP vendor had promised containerized options, but looks set to focus on the cloud

    ERP vendor Infor is to end development of on-premises and containerized versions of its core product for customers running on IBM iSeries mid-range systems.

    Born from a cross-breeding of ERP stalwarts Baan and Lawson, Infor was developing an on-premises containerized version of M3, dubbed CM3, to help ease migration for IBM hardware customers and offer them options other than lifting and shifting to the cloud.

    Infor said it would continue to run the database component on IBM i (Power and I operating system, formerly known as iSeries) while supporting the application component of the product in a Linux or Windows container on Kubernetes.

    Continue reading
  • Intel demos multi-wavelength laser array integrated on silicon wafer
    Next stop – on-chip optical interconnects?

    Intel is claiming a significant advancement in its photonics research with an eight-wavelength laser array that is integrated on a silicon wafer, marking another step on the road to on-chip optical interconnects.

    This development from Intel Labs will enable the production of an optical source with the required performance for future high-volume applications, the chip giant claimed. These include co-packaged optics, where the optical components are combined in the same chip package as other components such as network switch silicon, and optical interconnects between processors.

    According to Intel Labs, its demonstration laser array was built using the company's "300-millimetre silicon photonics manufacturing process," which is already used to make optical transceivers, paving the way for high-volume manufacturing in future. The eight-wavelength array uses distributed feedback (DFB) laser diodes, which apparently refers to the use of a periodically structured element or diffraction grating inside the laser to generate a single frequency output.

    Continue reading
  • Ex-Uber security chief accused of hushing database breach must face fraud charges
    Company execs and their lawyers are paying close attention to this one

    A US judge yesterday threw out an attempt to dismiss wire fraud charges against a former Uber employee accused of trying to cover up a computer crime.

    Former Uber security chief Joseph Sullivan is set to face criminal charges after US District Judge William Orrick yesterday [PDF] rejected his claim that prosecutors did not "adequately" allege that the goal of the claimed misrepresentation of the security breach was to get Uber's drivers to stay with the platform and continue paying service fees.

    In December last year, a federal grand jury handed down a superseding indictment adding wire fraud to the list of charges pending against Sullivan for his role in the alleged attempted cover-up of the 2016 security breach at Uber. The incident led to around 57 million user and driver records being stolen.

    Continue reading
  • FabricScape: Microsoft warns of vuln in Service Fabric
    Not trying to spin this as a Linux security hole, surely?

    Microsoft is flagging up a security hole in its Service Fabric technology when using containerized Linux workloads, and urged customers to upgrade their clusters to the most recent release.

    The flaw is tracked as CVE-2022-30137, an elevation-of-privilege vulnerability in Microsoft's Service Fabric. An attacker would need read/write access to the cluster as well as the ability to execute code within a Linux container granted access to the Service Fabric runtime in order to wreak havoc.

    Through a compromised container, for instance, a miscreant could gain control of the resource's host Service Fabric node and potentially the entire cluster.

    Continue reading

Biting the hand that feeds IT © 1998–2022