世界上最大规模的内存数据库SAP HANA产生

日前,Oracle总裁埃里森(Ellison),惹出不实言论,他在Oracle Open World上声称:" SAP HANA"规模太小,不及Exadata X3。声称HANA只有不到最大0.5TB的内存,而Exadata X3可以达到26TB。

这个不实言论,马上被SAP反驳,事实上,SAP HANA已经上线的系统,最大内存已经达到100TB,甚至可以扩至250TB。

真是冤家路窄。大规模内存数据库实时计算领域,完全看好HANA。

以下是完整的原文:

It could go down in IT history as the week Larry Ellison officially declared war on SAP HANA by making false claims about its own in-memory database product Exadata X3 and SAP’s platform for real-time business. Ellison, CEO of Oracle, is known for his brash public statements denigrating new trends that don’t fit into his world view, such as the Cloud and in-memory computing. After the remark, SAP wasted no time responding to Ellison’s statements in order to “set the record straight.”

Ellison’s assertions were made during his opening keynote speech at Oracle OpenWorld, the company’s annual customer event, held in San Francisco this week. Ellison claimed that the company’s Exadata X3 has 26 terabytes (TB) of memory, while purporting that SAP HANA is a “small machine” providing only 0.5 TB of memory.

“Clear distortions misrepresenting SAP HANA”

SAP Executive Board member Vishal Sikka: “We do not serve our stakeholders well by misstatements and omissions of key things we know to be true.”

SAP experts countered, saying that both claims were false. On Tuesday, Vishal Sikka, member of the SAP Executive Board and head of technology and innovation, posted a blog to the SAP sponsored website experiencesaphana.com expressing his concern about what he called Ellison’s “distortions.”

“Oracle made statements that are clear distortions and misrepresent SAP HANA,” wrote Vishal. “We do not serve our stakeholders well by misstatements and omissions of key things we know to be true.” Vishal then went on to compare the major differences between the two products, even inviting Ellison to Santa Clara, California, to visit the largest HANA system built so far. He was referring to an installation partnered with IBM. Steve Mills, senior vice president and group executive at IBM Software & Systems came to the support of SAP with a statement about the joint project: “IBM and SAP have partnered to demonstrate an SAP HANA system at 100 TB, making it the largest in-memory database system in the world. That system, running on 100 IBM nodes, can now scale to 250 TB.”

Steve Lucas, SAP’s general manager for Database & Technology, also refuted the claims made by Oracle’s CEO. Speaking to the web-based journal, Business Insider, Steve said: “Let’s just say it’s taken me 24 hours to get my eyebrows to lower to their normal position.” Steve provided his estimation of

SAP's advertizement in U.S. Newspapers in response to Oracle's false claims (click to enlarge).

Oracles Exadata X3: “Oracle is obviously following our lead, but their current efforts (X3) fall far short of what we are already delivering. X3 is still a hybrid system and an attempt to confuse and pacify customers.” Steve said that X3 consolidates databases onto one appliance and allows for some of the data to be stored in DRAM. “With a US$ 10 billion database albatross around Oracle’s neck, it’s obviously difficult for them to really to innovate,” he summed up Oracle’s current predicament.

1.     SAP HANA 系统可以是很小的系统也可以是超大规模的系统。SAP HANA的架构完全采用了大规模并行的多核系统,其全新设计的数据结构允许利用本地内存,这使得 SAP HANA 系统几乎拥有了无限制的扩展性。
2.     在过去几个月中,我们一直在供应由IBM、惠普、富士通和思科这四家厂商提供的、经认证的 16 节点 SAP HANA 硬件。这些系统均支持16TB的DRAM,所以已经比 Oracle的设备容量大4倍,并且从今年春季起已经在市场上供货了。在当前的配置下,这些设备最多能支持 32TB的DRAM。在 IBM的硬件上,配备 Max5 之后,还能支持到最多 40TB 的DRAM。
3.     迄今为止,世界上部署规模最大的 SAP HANA 系统是由英特尔和 IBM 于今年年初联合打造的 100 节点系统(见下图),它拥有 100TB 的 DRAM 以及 4000 个 CPU 内核。埃里森先生可以随时到加州的圣克拉拉市去亲眼看看,该系统可以处理若干客户的业务需求。在今年早些时候召开的SAPPHIRE NOW 大会上,我们与数以万计的客户分享了这一信息。如今,该系统已经能支持 250TB 的 DRAM(并且通过 HANA 压缩后,可以将几个PB的数据全部放在内存上。1PB=1024TB)。我们的合作伙伴 — IBM 软件和系统部高级副总裁兼集团执行官Steve Mills领导的团队帮助构建了该系统,并对这一开放式创新给予支持,他说:“IBM 和 SAP 联合打造了一款 100TB 的 SAP HANA 系统,并使之成为世界上最大的内存数据库系统。该系统在 100个 IBM X5节点配置下运行,现今可以扩展到250TB。”
拥有了我们合作伙伴提供的处理器和内存路线图后,这些系统的容量在今年秋季或2013年初的时候将会翻倍(因此,请把上边的数字乘以2)。同时我们不需要发布新硬件版本就可以获益于该创新成果。

SAP HANA构建于一个简单概念之上:随着硬件的发展和对现代企业软件本质的深入研究,使我们对数据库进行了重新思考。我们把这个概念变成了现实。我将这个概念视作 SAP HANA 构造的设计原则。大家都在尽力保护自己先前设计的数据库系统。新的 SSD 访问技术加快了对闪存的访问速度,提升了性能。SAP HANA 借助该技术,提升了读取日志和重启的速度,同时我们的 ASE 和 IQ 数据库也从中受益不少。但SAP HANA 的构架是以哈索博士多年以前的理论为基本原理:即当我们在内存中实施运行时,即便是最复杂的查询和算法,我们也能得到可预测的响应次数,而且任何程序都可通过海量并行程序同时执行。这赋予我们重新思考企业软件的自由空间。在不中断业务的情况下更新现有系统:从联机事务处理应用(例如我们上周在 SAP HANA 上发布的Business One 产品)到商务分析,从结构化数据处理到非结构化数据处理。重新构思系统,将运行速度提升数百万倍,并消除无所不在的批量作业。同样也让我们能够构建全新的应用和前所未有的解决方案。让软件简化并更好地、实时地连接我们的世界:从基因组排序到能源探测,从实时的客户交互到无所不能的银行业务。将我们从之前系统的局限中解放出来,使我们超越自己想象力的极限。
 
Alan Kay常对我说,未来不仅仅是历史的积累。我们选择聚焦未来,通过与客户与合作伙伴的合作,用我们的双手构建令人向往并且可以实现的未来,而不是一味地利用临时技术强化过时的、性能有限的系统。我认为,谎言将难以在未来立足。全世界都看到了基于SAP HANA的现状及其辉煌的成功,我们会继续打造其更美好的未来。
希望SAP HANA能带给您美好的体验!

节选自SAP 史维学博士(Vishal Sikka)

你可能感兴趣的:(oracle,数据库,IBM,database,System,SAP)