专栏首页technewsworld翻译专栏旧事物也可以是新事物

旧事物也可以是新事物

在过去的七个月或更长时间里,许多人的工作速度低于全速工作,但有些人却比其他人做得更多。即将发布影响CRM的大量新技术,但是即使没有到10月为止的最新公告,人们也意识到我们将全面发展。几十年前我们思考和争论的事情以不同的形式以新的解决方案返回。

Oracle再次使硬件变得多功能化,使用它来驱动新的业务模型并推销其CX版本的CRM。Salesforce正在解决其CRM产品线的一种大规模定制问题。

甲骨文

硬件一直很重要,但是一旦小型计算机可以与大型服务器建立网络,与运行大量用户相关联的许多焦虑就会消失。

即使用户在具有千兆字节可用的手持设备上,我们也不再刻苦思考用户需要多少千字节的内存。除非您正在运行很多用户,否则性能并不是问题。即使这对您来说是个问题,但向这个问题扔硬件比以往任何时候都更好。

但是更多的铁并不像哪种铁那么重要。当今的边缘应用程序需要大量的计算资源,因为它们会处理数据以提供推理引擎和算法,从而有助于实时决策,这是所谓的HPC或高性能计算的一部分。

所有这一切中的重要皱纹是如何从云中提供HPC -答案始于GPU或图形处理单元,今天相当于80287数学芯片。现代图形通过数学计算呈现,GPU处理数字。当然,GPU的另一种用途是将建议提供给面对客户的员工所需的所有概率处理。

本周,甲骨文宣布将在Oracle Cloud中全面支持Nvidia A100。该支持提供的性能可与Nvidia的DGX工作站相媲美,但通过云计算,Oracle表示“对于需要绝对性能来满足其工作负荷(如深度学习培训或硬件加速的可视化)的客户来说,这是一个很好的选择。”

翻译:您不再需要在大厅内托管计算密集型应用程序。云可以处理此问题,但此时您可以说运行自己的基准测试。

甲骨文公司此次发布的以及其他的硬件公告重点是为用户带来多种计算资源,而不是小型计算机时代所拥有的单一计算资源。但是,这也带来了大量的内存(在某些情况下高达2TB),因为没有什么比将数据从纳秒级的数据处理到纳秒要好得多了。这些都是我们数十年前所困扰的问题,现在仍然是问题,但是我们拥有更多工具。

在业务模型级别上的实际影响是消除了一些借口,这些借口没有将大型企业工作负载迁移到天堂。

销售队伍

退回机器的另一个问题是垂直市场CRM解决方案。销售通用CRM和提供针对行业最佳实践需求高度定制化解决方案的想法之间始终存在张力。换句话说,如何大规模定制CRM。

目标的移动主要是因为软件平台比以往任何时候都更容易构建可以针对垂直市场需求进行定制的通用功能。Salesforce最近以13亿美元收购的Vlocity联合创始人David Schmaier对此感到失望。尽管他认为需求少于24种行业应用程序,但他在Siebel负责监督。

Vlocity专门为少数几个大型行业提供垂直市场CRM应用程序。现在,借助Salesforce的资源,Vlocity应该能够做更多的事情,尤其是因为Salesforce已经具有行业定位。希望在未来几周内获得有关硬件加速和行业解决方案的更多公告。

我的两位

所有这一切都是在一个吉祥的时候出现的。已经需要可以交付给当今家庭战士的更好的应用程序和更多的计算能力。像这样的好消息在一年的这个时候像成熟的苹果一样到达,正好赶上了演出季节。今年只有虚拟表演,如果两家公司努力向前发展,Dreamforce和OpenWorld可以将Moscone中心变成超级传播的病毒孵化器-谁想要呢?

但是,这些公告以及其他类似公告为现在的需求打上了烙印。在过去的几个月中,我对1000多名面对客户的员工进行了两次调查,以了解他们的日常工作和技术使用情况。令我惊讶的是,它们大多数使用的系统是古董。

尽管在CRM上进行了大量投资,但我们实际上只是从头开始。甚至在过去几年中购买的系统都严重不足以满足当今的需求。我特别想起一项调查中的一项:CRM在人们用来进行电子邮件和社交媒体背后的面向客户的工作的应用程序列表中排名第四。

认真吗

更糟糕的是,面对客户的销售代表会为避免更新其CRM系统做很多工作。因此,似乎新一代支持用户的技术还远远不够。GPU处理器和行业特定的CRM等技术的进步将为我们的日常工作增加力量。

原文题目:What's Old Is New Again

原文:Many people have worked at less than full speed over the last seven months or more, but some people managed to get more done than others. A raft of new technologies that impact CRM are about to be announced, but even without the latest announcements due now through October, there's a realization that we're coming full circle. Things we thought about and argued over decades ago are back, in different forms, with new solutions.

Oracle is making hardware sexy again, using it to drive new business models and to push its CX version of CRM; and Salesforce is tackling a kind of mass customization of its CRM product line.

Oracle

Hardware has always been important, but once small computers could be networked with larger servers, much of the angst associated with running large groups of users melted away.

We no longer think very hard about how many kilobytes of memory a user needs, even on handheld devices that have gigabytes available. Performance isn't much of an issue unless you are running many, many users banging on an inference engine. Even if that's an issue for you, throwing hardware at the problem works better than ever.

But more iron isn't as important as which iron. Today's marginal apps are compute hungry because they crunch data to feed inference engines and algorithms that assist in real time decision-making, part of what's called HPC or High Performance Computing.

The important wrinkle in all this is how to deliver HPC from the cloud -- and the answer starts with the GPU or Graphic Processing Unit, today's equivalent of the 80287 math chip. Modern graphics are rendered through math calculations and the GPU crunches the numbers. Naturally another use for the GPU is all of the probability processing needed to deliver recommendations to customer-facing employees.

This week Oracle announced general availability of Nvidia A100 support in the Oracle Cloud. This support provides performance comparable to Nvidia's DGX workstation but through the cloud which Oracle says is "A great alternative for customers that need the absolute performance for their workloads such as deep learning training or hardware accelerated visualizations."

Translation: You don't need to host your compute intensive apps down the hall any more. The cloud can handle that, but at this point it's good to say run your own benchmarks.

This and the rest of Oracle's hardware announcements focus on bringing multiple compute resources to users rather than the single ones available back in the minicomputer era. But this also brings a good deal of memory -- up to 2TB in some cases -- because there's nothing like having data available to be crunched in nanoseconds instead of waiting milli seconds to get the data from disk. These were all issues we obsessed over decades ago, and they're still issues, but we have many more tools.

The practical impact at the business model level is to remove a layer of excuses for not migrating even large corporate workloads to the heavens.

Salesforce

Another issue from the way back machine, is vertical market CRM solutions. There's always been a tension between the idea of selling generic CRM and that of delivering solutions highly customized to the needs of an industry's best practices. In other words, how to mass customize CRM.

The target has moved primarily because software platforms make it easier than ever to build a generic function that can be customized to vertical market needs. David Schmaier, co-founder of Vlocity, a company that Salesforce paid US$1.3 billion for recently, has this down. Though he sees a need for fewer than the 24 industry apps he oversaw at Siebel.

Vlocity specializes in vertical market CRM apps for a handful of large industries. Now, with Salesforce's resources behind it, Vlocity should be able to do more, especially since Salesforce already had an industries orientation. Look for more announcements on hardware acceleration and industry solutions in the weeks ahead.

My Two Bits

All of this comes along at an auspicious time. There's already a need for better apps and more compute power that can be delivered to today's home warriors. Propitious announcements like these arrive at this time of year like ripe apples, in time for show season. There will only be virtual shows this year, Dreamforce and OpenWorld could turn the Moscone Center into a super spreading virus incubator if the companies tried to go forward -- and who wants that?

But these announcements and others like them hit the mark for what's needed now. Over the last few months I conducted two surveys of over one thousand customer-facing employees to learn about their daily routines and their use of technology. To my surprise, the systems most of them use are antiques.

Despite the big investments in CRM we've really only scratched the surface. Systems bought even in the last few years are woefully inadequate for today's demand. One item from one survey especially sticks in my mind: CRM was fourth from the top of the list of apps that people rely on to do customer-facing jobs behind email and social media.

Seriously?

Worse, customer-facing reps would do much to avoid updating their CRM systems, and do. So it looks like a new generation of technology to support users can't come soon enough. Advances in things like GPU processors and industry specific CRM will add muscle to our day to day work effort.

原文作者:By Denis Pombriant

原文链接:https://www.technewsworld.com/story/86858.html

翻译人:田冠宇

原文链接:https://www.technewsworld.com/story/86858.html

原文作者:By Denis Pombriant

我来说两句

0 条评论
登录 后参与评论

相关文章

  • 使用循环强化学习和模拟退火将其放置在集成电路中(AI LG)

    随着IC技术的复杂性不断提高,集成电路(IC)的物理设计和生产正变得越来越具有挑战性。布局一直是IC物理设计中最关键的步骤之一。通过数十年的研究,基于分区,以及...

    田冠宇
  • AntiX Linux:外观不漂亮,但功能强大

    该antiX发行版提供了坚实的操作系统,许多其它的Linux社区不能提供。它虽然张扬,但可靠,并且专注于性能而不是美观。

    田冠宇
  • 为什么要为AI,IoT和ML采购外国人才

    人工智能革命已经进行了一段时间,但这并不意味着所有软件研发都属于其中。尽管如此,它还是推动软件行业前进的有趣例子。

    田冠宇
  • Work.com与社会的客户关系管理化(Internet)

    一个前一段时间我记录与和未来的CRM公司播客有点集中在银行和信用社,CRMNext,并提出了涉及什么是对CRM地平线的问题之一。老实说,没有人真正知道,但是因为...

    用户8054058
  • 六个也许你需要转变CRM的标志(computing application)

    在当今不断发展的业务环境中,每个运营决策都是至关重要的-包括管理客户的最佳实践。客户关系管理平台是该过程的组成部分。实际上,拥有11名以上员工的公司中有91%使...

    木樾233
  • 为2021年的CRM提速

    到了2000年,有了托管产品,成为软件即服务,最终成为云计算。SaaS非常流行,因为它解决了购买和实施一个大系统的经济问题。我们刚刚做了千年变化的ERP,许多人...

    用户8128510
  • 无崩溃,无漏洞:嵌入式内核的自动验证(CS)

    内核是许多计算机系统中最安全和最关键的组件,因为最严重的错误会导致整个系统崩溃或被利用。因此,使用正式的方法来保证内核不受这些错误的影响是可取的,但是这样做所需...

    柴艺
  • 原创译文 | 区块链技术将引入游戏中!期待吗?

    导读:上一期了解了关于有感知力的AI的相关介绍,今天我们来了解一下区块链技术应用到游戏的相关内容(文末更多往期译文推荐) 在以往游戏过程中,购买游戏内物品的概念...

    灯塔大数据
  • [技术报告]将抽样和梗概与最坏情况下最佳运行时间和质量保证用于图模式基数估计(CS)

    图模式基数估计是估计查询图在数据图中嵌入的数量。这个基本问题就产生在,比如说,子图匹配算法中的查询规划。解决这个问题有两种主要方法:采用和概要。如果概要能很好地...

    用户8440711
  • 【论文推荐】最新七篇知识图谱相关论文—知识表示学习、增强神经网络、链接预测、关系预测与提取、综述、递归特性生成、深度知识感知网络

    【导读】专知内容组整理了最近七篇知识图谱(Knowledge graphs)相关文章,为大家进行介绍,欢迎查看! 1. Does William Shakesp...

    WZEARW

扫码关注云+社区

领取腾讯云代金券