Zynga宣布关闭花费1亿美元构建的数据中心,仍回归AWS

游戏公司Zynga决定关闭自建数据中心,转向使用亚马逊AWS云服务,这一转变背后的原因在于其业务特性使得云服务更加适合其需求。Zynga曾花费1亿美元建设数据中心,但发现自建数据中心成本高昂且难以应对游戏行业的流量波动和不可预测的增长。最终,Zynga选择让AWS承担数据中心工作,以更灵活地应对业务变化。
部署运行你感兴趣的模型镜像

对于一家游戏公司而言,Zynga的故事说明最符合企业利益发展的是研发新游戏,而不是在投入巨大,运维复杂的数据中心上占用太多精力。

原文地址:http://www.datacenterknowledge.com/archives/2015/05/12/zynga-ditches-data-center-plans-aws/

Casual and social gaming provider Zynga will shut its data centers and shift workloads back to Amazon, two years after the company spent $100 million to build out its own data centers, according to The Wall Street Journal.

Zynga appeared to be a business that could potentially “outgrow” cloud, in the sense that once a company reaches a certain scale, it begins to make sense to build your own data centers.

However, running your own data center is capital intensive. It was the nature of Zynga’s business that made this a tricky proposition. Zynga didn’t grow as anticipated, following bursts of growth. This is particularly true in regards to its growth outside of the Facebook ecosystem.

Online gaming is subject to unpredictable traffic spikes and drastic sways of traffic dependent on the release schedule. It’s often feast or famine for a game publisher, given the unpredictable nature of what gamers glom on to, and the often short lifespan of a title. The type of unpredictable workloads and growth Zynga faced are tailor-made for cloud.

For Zynga to run its own data centers, it meant a mad scramble to provision servers and power during big releases, followed by lulls. This kind of wide variation is hard to accomplish on your own servers.

Zynga CEO Mark Pincus said on an earnings call that there a lot of strategic places for them to have scale, but running their own data centers isn’t one that’s appropriate. “We’re going to let Amazon do that,” he said.

The company has always relied on AWS for some workloads, even at the height of using its own data centers. Zynga created software called zCloud that helped it shift workloads between Amazon’s servers and its own data center.

Zynga disclosed its intentions to boost data center investment when it filed for its Initial Public Offering (IPO) in 2011. The company said it could operate its own data centers more cheaply than Amazon. Amazon’s economies of scale proved hard to beat.

Zynga leased data center space from two wholesale data center providers, DuPont Fabros Technology and Digital Realty Trust. Another indication of the company’s close roots to Facebook – several of the facilities were adjacent to Facebook’s data center facilities.

Zynga’s rode social media’s coattails to build a casual gaming empire with initial hits like FarmVille and Mafia Wars. Heavily dependent on Facebook users, Zynga suffered whenever Facebook made changes to the feed unfavorable to gaming content.

您可能感兴趣的与本文相关的镜像

ACE-Step

ACE-Step

音乐合成
ACE-Step

ACE-Step是由中国团队阶跃星辰(StepFun)与ACE Studio联手打造的开源音乐生成模型。 它拥有3.5B参数量,支持快速高质量生成、强可控性和易于拓展的特点。 最厉害的是,它可以生成多种语言的歌曲,包括但不限于中文、英文、日文等19种语言

基于可靠性评估序贯蒙特卡洛模拟法的配电网可靠性评估研究(Matlab代码实现)内容概要:本文围绕“基于可靠性评估序贯蒙特卡洛模拟法的配电网可靠性评估研究”,介绍了利用Matlab代码实现配电网可靠性的仿真分析方法。重点采用序贯蒙特卡洛模拟法对配电网进行长时间段的状态抽样与统计,通过模拟系统元件的故障与修复过程,评估配电网的关键可靠性指标,如系统停电频率、停电持续时间、负荷点可靠性等。该方法能够有效处理复杂网络结构与设备时序特性,提升评估精度,适用于含分布式电源、电动汽车等新型负荷接入的现代配电网。文中提供了完整的Matlab实现代码与案例分析,便于复现和扩展应用。; 适合人群:具备电力系统基础知识和Matlab编程能力的高校研究生、科研人员及电力行业技术人员,尤其适合从事配电网规划、运行与可靠性分析相关工作的人员; 使用场景及目标:①掌握序贯蒙特卡洛模拟法在电力系统可靠性评估中的基本原理与实现流程;②学习如何通过Matlab构建配电网仿真模型并进行状态转移模拟;③应用于含新能源接入的复杂配电网可靠性定量评估与优化设计; 阅读建议:建议结合文中提供的Matlab代码逐段调试运行,理解状态抽样、故障判断、修复逻辑及指标统计的具体实现方式,同时可扩展至不同网络结构或加入更多不确定性因素进行深化研究。
评论
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值