The State of Software-Defined Storage, Hyperconverged and Cloud Storage

本文链接：https://blog.csdn.net/u014454538/article/details/80282460

#The State of Software-Defined Storage, Hyperconverged
and Cloud Storage—SIXTH ANNUAL MARKET SURVEY

In the 2017 Strategic Roadmap for Compute Infrastructure, analyst firm Gartner states that “compute infrastructure is rapidly evolving from hardware-centric silos to software-driven ecosystems where application agility drives infrastructure architectures. I&O (infrastructure and operations) leaders must plan now to address these trends impacting computing to evolve and modernize their infrastructure for digital business.1”

在2017年计算基础架构战略路线图中，分析公司Gartner指出：“计算基础架构正在迅速从以硬件为中心的筒仓向由应用灵活性驱动基础架构架构的软件驱动的生态系统演变而来。 I＆O（基础架构和运营）领导者现在必须计划应对这些影响计算的趋势，以便为数字业务发展和现代化基础架构.“

For the sixth consecutive year, the DataCore Software annual survey explores the impact of major software-driven storage deployments within organizations across the globe. The survey distills the expectations and experiences of 426 IT professionals who are currently using or evaluating software-defined storage (SDS), hyperconverged and cloud storage to solve critical data storage challenges. The results yield surprising insights from a cross-section of industries over a wide range of workloads.

##生词、短语

生词/短语	读音	释义
consecutive	`[kən'sekjʊtɪv]`	adj. 连贯的；连续不断的
silo	`['saɪləʊ]`	n. 筒仓；青贮窖；飞弹发射井 vt. 把…存入青贮窖；把…储存在筒仓内
ecosystem	`['iːkəʊsɪstəm]`	n. 生态系统
agility	`[ə'dʒɪlətɪ]`	n. 敏捷；灵活；机敏
distill	`[dis'til]`	vt. 提取；蒸馏；使滴下 vi. 蒸馏；滴下；作为精华产生（等于distil）
hyperconverged	`[kən'vɜ:dʒd]`	超融合
yield	`[jiːld]`	n. 产量；收益 vt. 屈服；出产，产生；放弃 vi. 屈服，投降
demographics	`[demə'græfɪks]`	n. 人口统计资料
respondent	`[rɪ'spɒnd(ə)nt]`	n. [法] 被告；应答者 adj. 回答的；应答的
future-proof		n. 不会过时的技术
lock-in	`['lɔkin]`	n. 禁闭；占据；锁入
portray	`[pɔː'treɪ]`	vt. 描绘；扮演
CAPEX		capital expenditures，abbr. 资本支出
versus	`['vɜːsəs]`	prep. 对；与…相对；对抗
OPEX		Operating Expense,abbr. 运营成本；运营支出；运营开支
exponential	`[,ekspə'nenʃ(ə)l]`	n. 指数 adj. 指数的
disaster	`[dɪ'zɑːstə]`	n. 灾难，灾祸；不幸
ballpark	`['bɔːlpɑːk]`	n. （美）棒球场；活动领域；可变通范围 adj. 大约的
hypervisor	`[,haipə'vaizə]`	n. 超级监督者；管理程序
combat	`['kɒmbæt; 'kʌm-]`	n. 战斗；争论 adj. 战斗的；为…斗争的 vi. 战斗；搏斗 vt. 反对；与…战斗
off-the-shelf	`['ɔfðə'ʃelf, 'ɔ:-]`	adj. 现成的；常备的；成品的 adv. 现成地；无需作重大修改地
incentive	`[ɪn'sentɪv]`	n. 动机；刺激 adj. 激励的；刺激的
ISV		independent software vendor,abbr. 独立软体开发商
prevalent	`['prev(ə)l(ə)nt]`	adj. 流行的；普遍的，广传的
spinning	`['spɪnɪŋ]`	n. 纺纱 v. 旋转，纺织（spin的现在分词） adj. 纺织的
velocity	`[vəˈlɒsəti]`	超融合
adequately	`['ædikwitli]`	adv. 充分地；足够地；适当地
agnostic	`[æg'nɒstɪk]`	n. 不可知论者 adj. 不可知论的
indisputably	`[,indi'spju:təbli]`	adv. 无争论余地地
mission-critical		关键任务
bandwagon	`['bændwægən]`	n. 流行，时尚；乐队花车
rule out		排除；取消；划去；反对；阻止
hesitancy	`['hezɪt(ə)nsɪ]`	n. 踌躇，犹豫
regulatory	`['reɡjulətəri]`	adj. 管理的；控制的；调整的
counterpart	`['kaʊntəpɑːt]`	n. 副本；配对物；极相似的人或物
coexistence	`[,kəuiɡ'zistəns]`	n. 共存；和平共处；并立

##基础知识

What are the business drivers for implementing software-defined storage?

“I&O leaders are looking for softwaredefined storage (SDS) products that offer the potential for better total cost of ownership (TCO), efficiency and scalability to address exponential-data growth needs, and to benefit from innovations from hardware and software players independently.”

“I＆O（infrastructure and operations）领导者正在寻找能够提高总体拥有成本（TCO）、效率和可扩展性的软件定义存储（SDS）产品，以解决指数数据增长需求，并从硬件和软件的独立创新中受益。”

“What are the primary capabilities that you would like from your storage infrastructure when virtualizing storage?”

The majority of respondents (83%) replied that business continuity from high availability was the top concern (metro clustering, synchronous mirroring), while more than two thirds (73%) replied that enabling storage capacity expansion without disruption was a primary capability of importance. Cost efficiency and disaster recovery (asynchronous replication to remote site) also ranked high, coming in at 65% and 60% respectively. These top two priorities were also the same in past surveys, with cost reductions for greater infrastructure lifetime savings coming in third in past years.

这里写图片描述

overcome performance problems

The need for faster databases and data analytics is driving new requirements for technologies to optimize performance and meet demand for real-time responses. To address these performance problems, the top approaches that respondents identified as preferred methods to overcome performance problems include using all-flash arrays; software acceleration on the host machine and switching to in-memory databases.

Which of the following approaches do you prefer to overcome performance problems, ranked in preferred order with 1 as most desirable, 5 least desirable?
这里写图片描述

It is clear that flash technologies are now more prevalent and have helped solve many performance issues by replacing spinning disks with electronic media. From a media standpoint, this is goodness, and also brings along other cost-saving aspects such as smaller footprints and power requirements. However, the primary performance problem of processing I/O for high-velocity transactions, write traffic and latency-sensitive applications like databases has not been adequately addressed by technologies such as all-flash arrays and NVMe protocols. They have helped, but at the same time have added additional expenses and still don’t fully take advantage of multiprocessing capabilities.

This again ties to the problem that most computing processes work in serial, underutilizing the multi-core capabilities inherent in today’s off-the-shelf server platforms. While in-memory technology has become “hot” for database applications (and while it’s important to take advantage of faster memory speeds) once again, I/O processing and faster response aren’t adequately addressed by this.

This illustrates that flash devices are beginning to expand from a limited role inside servers to wider use within the storage infrastructure.这说明闪存设备正在从服务器内的有限角色开始扩展到存储基础架构内的更广泛使用。

Current State of Hyperconverged Infrastructure

① software-centric architectures that integrate compute, storage and networking on commodity hardware. （Gartner’s definition）
以软件为中心的架构，在商用硬件上集成计算、存储和网络。
② hyperconverged systems collapse core storage and compute functionality into a single, highly virtualized solution. A key characteristic of hyperconverged systems that differentiate these solutions from other integrated systems is their ability to provide all compute and storage functions through the same server-based resources.（IDC）
超融合系统将核心存储和计算功能折叠成一个高度虚拟化的解决方案。将这些解决方案与其他集成系统区分开来的超融合系统的一个关键特征是它们能够通过相同的基于服务器的资源提供所有计算和存储功能。
③ An approach to technology infrastructure that packages server, storage, and network functions into a modular unit and adds a software layer to discover, pool, and reconfigure assets across multiple units quickly and easily without the need for deep technology skills. These systems can be implemented either as software plus modular physical units or as a software overlay on top of existing infrastructure.（Forrester）
一种将服务器、存储和网络功能打包为模块化单元并添加软件层的技术基础架构方法，可快速轻松地在多个单元中发现、汇集和重新配置资产，而无需深入的技术技能。这些系统既可以作为软件加模块化物理单元，也可以作为现有基础设施之上的软件覆盖层来实现。

From additional customer discussions on the subject of hyperconverged adoption, while many hyperconverged systems work well for simple use cases such as a remote office or with virtual desktop infrastructures — where only a few nodes are needed — it’s not always suitable for more challenging cases that require “real-time” response or depend on input-output(I/O)-intensive workloads like databases. Also, the same discussions revealed that while hyperconverged was a good solution for these simpler use cases, the lack of flexibility and the inability to work with existing storage and infrastructure also created hyperconverged ‘silos’ which could not be easily managed or integrated within the overall company infrastructure.

从超融合采用方面的其他客户讨论中，尽管许多超融合系统适用于简单用例（如远程办公室或虚拟桌面基础架构 - 仅需要少量节点），但并非总是适合需要更具挑战性的情况 “实时”响应还是依赖于输入输出（I / O）密集型工作负载（如数据库）。此外，同样的讨论表明，虽然超融合对于这些较简单的用例来说是一个很好的解决方案，但缺乏灵活性和无法与现有存储和基础架构一起工作也会产生超融合的“孤岛”，无法在整体内轻松进行管理或整合公司基础设施。