b********1
发帖数: 137 | 1 Presented by engineering teams from Microsoft, Google, Facebook, Uber,
Airbnb, LinkedIn, Pinterest, Twitter, etc.., Join us at the 3rd NEXTCON
, the leading conference for AI, cloud and data technology, taking place on
Sep 15-16th in Seattle. It brings together top technical engineers,
practitioners, influential technologist and data scientists to share their
solutions and practical experience in tackling the pressing problems. So we
can learn lessons and best practices from experts, exch... 阅读全帖 |
|
h******z
发帖数: 32 | 2 Hiring Sr SDE in Expedia. Sponsor H1-B and GC.
Position summary
Are you passionate about data and the technology needed to drive a multi-
billion dollar business? Do you love building creative, high-scale data
products using a diverse set of state of the art technologies?
We are looking for a Software Development Engineer with a real passion for
data. We are responsible for building accelerators, products and processes
that power the data pipelines at Expedia, enable self-service capabilities
an... 阅读全帖 |
|
d***n
发帖数: 39 | 3 网络个人理财公司, 在DC downtown.
Job Description:
Description:
We seek solid Java developers who do not shy away from challenges. We’re a
fun and dynamic startup, with a team of people who work together on
everything from the latest AJAX front-end technologies to superb database
design. If you join our team, you’ll work in our agile engineering group,
contributing your expertise to all aspects of the software.
Required Skills/Experience
3-5 years experience developing commercial quality Web applicati... 阅读全帖 |
|
N***M
发帖数: 4295 | 4 Salary 120k-130k+bonus - Senior Java developer position in Sterling/Ashburn
VA
Duration Full time Permanent
Start 1-2 weeks
Location Sterling VA/ Ashburn VA - near Dulles Mall
The Senior Software Engineer works as part of a small, highly motivated
software/hardware development team creating application solutions for an
education media product line that captures the classroom experience (audio/
video/display) and delivers it as streaming media over the internet. This
person is responsible for res... 阅读全帖 |
|
k********n
发帖数: 1819 | 5 不懂还冲大头，收入用中间值，什么时候消费也用中值？?你先说怎么统计吧。
一般scalable的考虑分布，用中值，不scalable的就还是平均值，比较可靠，也易于统计 |
|
p****s
发帖数: 3184 | 6 是社会渐渐脱离小农社会，越来越复杂，复杂度越来越高的体现。
人治模式在这种复杂社会里是忙不过来的，人治就是个主从服务器的小规模分布式系统
，不scalable。要scalable，必须得像Internet那样搞法治(协议)系统。 |
|
d****y
发帖数: 472 | 7 Please send your resume to a************[email protected]
Senior Software Engineer
Responsibilities:
· Create new features from scratch, enhance existing features and
optimize existing functionality, from conception and design through testing
and deployment
· Architect and develop new software products using C/C++ in
Windows and Unix/Linux environment
· Work closely with product management to ensure we are building
the best products
· Work with engineering support team ... 阅读全帖 |
|
ET
发帖数: 10701 | 8 http://ieeexplore.ieee.org/Xplore/login.jsp?url=/iel5/9427/29913/01365763.pdf
Generation and integration of scalable bipolar compact models [HBT example]
Sheridan, D.C.; Murty, R.M.; Newton, K.M.; Johnson, J.B.
Bipolar/BiCMOS Circuits and Technology, 2004. Proceedings of the 2004
Meeting
Volume , Issue , 13-14 Sept. 2004 Page(s): 132 - 139
Digital Object Identifier 10.1109/BIPOL.2004.1365763
Summary: This paper investigates the methodologies for generating and
integrating scalable bipolar mode |
|
b***i
发帖数: 3043 | 9 我就是想看看真正scalable的spring的例子，就是有scalable的数据存储，存对象，比
如图片，音乐什么的。看看怎么回事，调试一下，改改代码，了解了解。
问题是，Openshift不是还卖gear吗？难道还同时卖服务做gear上的应用? |
|
v******d
发帖数: 1322 | 10 我想cloud应该是用来解决scalability的问题，在某些目标上跟grid有点像，不过因为
是针对的应用不一样，设计有很大差别，还有结合了virtualization, software/
platform/infrastructure as a service等等的概念和技术。
"初期费用低”可以理解为cloud提供了一些解决scalability的hardware和software
layers, 只要在这上面开发自己的程序就可以了，这样还是费用还是能够降低的。如
果是小型应用，应该降低不了多少初期费用。
还有维护一个大型系统费用还是很高的，这点不仅仅是“整点STORAGE，装装SP，整两
人看着有没有流量异常...”。 |
|
B*****g
发帖数: 34098 | 11 【 以下文字转载自 Java 讨论区 】
发信人: verde (没有你成), 信区: Java
标 题: Re: Job with Oracle PL?
发信站: BBS 未名空间站 (Wed Dec 21 06:21:14 2011, 美东)
scalability这个词是有context的，不同的context不能直接比较，
RDBMS和NoSQL是一个spectrum里面两个不同的position，直接
比scalability好比apple vs orange，恐怕不太合适。
选nosql还是rdbms完全是application domain的问题，和这两种技术
谁高谁低无关。
PL/SQL完全可以run 在cluster上，我不知道前面说pl/sql只能run在一个
machine上是怎么来的， |
|
f*****m
发帖数: 416 | 12 IPv6为了scale, 路由要一路aggregate,这样核心路由表才能小.这样客户的路由就是
provider dependent的.问题是当年讨论这个的时候,只注意了scalability,没注意到客
户multi-homing的需要.客户如果multi-homing的话,就要多个PD的地址段,每台host也
会有多个unicast address (PD). 多个PD又带来诸多问题, load sharing, 换SP的时候
renumbering...Shim6历经多年才变成RFC,不过修成后也没谁用.
后来有开始用PI的地址段,这样换provider的时候不用renumber, 可是PI的后果是
scalability被破坏. 再后来LISP出来了,不过持怀疑态度的一大堆.再后来就是这个了.
.. |
|
s*****g
发帖数: 1055 | 13 Is this not a simple MPLS-VPN hub-spoke configuration? or if you are
concerned by VRF configuration scalability, use scalable MPLS-VPN hub-spoke
configuration by using upstream/downstream VRFs? |
|
s*****g
发帖数: 1055 | 14 Yes, but if PE1 has two or more CEs attached, then the two CEs will be able
to communicate with each other without going to hub unless you use different
VRF, which is not scalable. You need "scalable MPLS-VPN hub-spoke" feature
to circumvent this limitation. |
|
k********h
发帖数: 22 | 15 About Us:
ROBLOX is a leading user-generated gaming site that makes players the
architects of their own 3D worlds, drawing over one billion page views and
21 million in-game hours each month. The site is ranked #1 in the US for
total engagement time within the 8-14 year old segment, and is rapidly
growing in popularity amongst players aged 15+. Our vision is to become the
leader in user-generated gaming for all ages, around the world. On ROBLOX,
players use core building components to create ela... 阅读全帖 |
|
z**r
发帖数: 17771 | 16 这俩没有冲突吧？一个是解决L2 CONVERGENCE加ECMP加MAC SCALABILITY，另外一个解
决VLAN SCALABILITY |
|
L******t
发帖数: 1985 | 17 The main reason to use L2 DCN is because of VM mobility. Now with VXLAN, VM
mobility is supported as well, but this time over L3 DCN. In L3 DCN,
multipath is no longer a problem. And VXLAN can also fix the MAC scalability
issue in similar way as FabricPath. And VXLAN is even better because it
improves vlan scalability.
So in my understanding VXLAN makes FabricPath/TRILL much less appealing than
before. |
|
s******v
发帖数: 4495 | 18 总结
1. Netflix Open Connect是NFLX自己基于开源和off-shelf硬件而开发的一个CDN系统
2. 有两种CDN实现模式，
1) ISP从NFLX拿到OCA，建立BGP链接，作为Local Stream Server，具体的看下面
control flow;
2) NFLX在大的Internet Exchange里面建立Local CDN cluster, 和ISP直接链接
NFLX占了非常高的Internet traffic比例, 2H 2013 vs 2H 2012
1) NFLX 32% ~ 33%
2) Youtube 18% ~ 14%
Netflix decrease illegal file sharing.
项目的背景：
1）项目从2011年开始，目前已经offload了100%的Traffic，达到了16TB；而其他的CDN
例如，Akamai，Limenetwork，Level 3只能达到20-25%；
2）4个Racks的数据量可以达到8.6PB
3）multiple-tier arch:
3.1) Headend, 1(6?) OCA... 阅读全帖 |
|
s*****g
发帖数: 1055 | 19 Spine-leaf 显然是DC不二的选择，cheap, flexible, scalable, no vendor lock in.
至于L2 还是L3 还要看你application, scalability 的需求，如果只有少量的rack,
spanning tree works just fine, L3 to TOR 应该是标配，需要L2 inside D.C. 话可
以用evpn/vxlan 实现，trill/qfabric 应该不是大势所趋。 |
|
s*****g
发帖数: 1055 | 20 Spine-leaf 显然是DC不二的选择，cheap, flexible, scalable, no vendor lock in.
至于L2 还是L3 还要看你application, scalability 的需求，如果只有少量的rack,
spanning tree works just fine, L3 to TOR 应该是标配，需要L2 inside D.C. 话可
以用evpn/vxlan 实现，trill/qfabric 应该不是大势所趋。
Buffer size? there are two totally opposite opinions wrt why deep buffer
matter/does not matter ... it is beyond my pay grade. |
|
发帖数: 1 | 21 Intel日前发布了Skylake-SP平台的Xeon Scalable处理器家族，也就是E5/E7改名，现
在分别按照性能高低划分为铜、银、金、铂金四档。
牙膏挤多了？Intel新Xeon处理器性能暴力提升59%
Intel对外公布了Xeon Scalable的性能参数，称在SAP HANA数据库内存计算应用中，新
一代Xeon快了1.59倍。 |
|
q***e
发帖数: 90 | 22 To be simple, whether you want 3-tier architecture or 2-tier
architecture?
Without an application server, using JDBC etc directly
skipped the EJB part. Two tier definitely works faster than
3 tier, but again you lose the scalability.
If you don't care distributed processing and scalability,
3-tier does not have much advantage. In terms of maintanence,
3-tier might be better, since change in front and backend are
seperated. Developing time and debuging time really depends on
th |
|
d***n
发帖数: 39 | 23 网络个人理财公司, 在DC downtown.
Job Description:
Description:
We seek solid Java developers who do not shy away from challenges. We’re a
fun and dynamic startup, with a team of people who work together on
everything from the latest AJAX front-end technologies to superb database
design. If you join our team, you’ll work in our agile engineering group,
contributing your expertise to all aspects of the software.
Required Skills/Experience
3-5 years experience developing commercial quality Web applicati... 阅读全帖 |
|
v***e
发帖数: 2108 | 24 scalability这个词是有context的，不同的context不能直接比较，
RDBMS和NoSQL是一个spectrum里面两个不同的position，直接
比scalability好比apple vs orange，恐怕不太合适。
选nosql还是rdbms完全是application domain的问题，和这两种技术
谁高谁低无关。
PL/SQL完全可以run 在cluster上，我不知道前面说pl/sql只能run在一个
machine上是怎么来的， |
|
g*****g
发帖数: 34805 | 25 Oh man, don't get me started again. In 2007, I said SL is a waste,
you said SL is the best thing since slice bread. And see what SL
is at right now.
And don't talk about what can Java counter .Net. .Net is the catchup
guy, remember? You should ask what .Net has really replaced other
than the old VB apps? Last time I check, IBM and Oracle are still
the brand names in enterprise market, not Microsoft. The social networks
are still running on Linux. FB started with the cheap LAMP and moves
part of ... 阅读全帖 |
|
S*******h
发帖数: 7021 | 26 【 以下文字转载自 JobHunting 讨论区 】
发信人: StanMarsh (Stan), 信区: JobHunting
标 题: 多個Software Engineer 職缺
发信站: BBS 未名空间站 (Mon Sep 10 20:35:35 2012, 美东)
Multiple software developer and software engineer in test openings. Send me
a message if you are interested in.
Software Engineer Description
Would you like to use your expertise to help create exciting new cloud-based
products and services? The successful candidate will be working with a
small engineering team to design and build high performance, scalable ... 阅读全帖 |
|
z*******3
发帖数: 13709 | 27 我这点很同意tom说的
一旦你用了，你就爱不释手
不会之前你会觉得很可怕
但其实没有那么可怕
无非是个工具，你只要理解了这些东西做什么用的
剩下的就很简单，无非做两个试验，照葫芦画瓢
先看定义，javaee的定义，摘取自wikipedia
Java EE includes several API specifications, such as JDBC, RMI, e-mail, JMS,
web services, XML, etc., and defines how to coordinate them. Java EE also
features some specifications unique to Java EE for components. These include
Enterprise JavaBeans, Connectors, servlets, JavaServer Pages and several
web service technologies. This allows developers to create enterprise
applicati... 阅读全帖 |
|
g*****g
发帖数: 34805 | 28 其实做大的Android app一多半都有后台，而且稍微知名的用户数目就是几个million。
对后台服务的scalability要求可能很高。startup又没钱搞自己的data center，最常
见就是AWS。所以mobile, cloud, scalability，所有的热点词都整到一块了。看着合
适再上点NoSQL DB，用Hadoop 做点数据分析，基本上天天LinkedIn得有人来骚扰。去
锻炼一下运气好就
instagram了，运气不好跳槽也很容易。 |
|
t*********e
发帖数: 630 | 29 1. Cassandra
The Cassandra database serves as a "scalable system of record" in the big
data world, says Jonathan Ellis, vice president of the Cassandra project.
Apache received the project from Facebook, which open-sourced Cassandra in
2008. Whereas Hadoop undertakes data analysis, Cassandra provides a data
store for applications, often highly scalable ones on the Web. Netflix, for
example, runs many Cassandra clusters, Ellis says.
2. Cordova
Giving Apache prominence in mobile computing, Cordova... 阅读全帖 |
|
j********r
发帖数: 21 | 30 What's kind of scalability you are talking? There are many ways to deal with
scalability. |
|
T*******x
发帖数: 8565 | 31
哦。“填加硬件而不需要修改软件”的能力？知道了。谢谢。
horizontal scalability是指可以添加node，比如服务器等。
vertical scalability是指可以在同一台机器上添加比如内存什么的。 |
|
h*******e
发帖数: 225 | 32 不知道你老是说C++有scalability问题是什么意思，是说coding上不容易写大系统，
还是说做出来的系统不scalable. |
|
g*****g
发帖数: 34805 | 33 You worry too much. Scalability is not an issue until
it's an issue. And for average website, it may never
be an issue.
ZK is one of so called stateful component frameworks, components
are rendered on server side. But traditional request-based
frameworks are done this way too. The main advantage of ajax,
is updating only a portion of webpage so that the update can
be fast. Reducing IO isn't as important. That being said, as
long as it's ajax framework, you can still expect less traffic
than requ... 阅读全帖 |
|
g*****g
发帖数: 34805 | 34 That's a big topic. But if you ever heard of CAP theorem.
Basically it's availability vs. consistency.
Traditional DB is consistent but cannot be clustered in
linear scalability. NoSQL DB uses so called eventual
consistency to achieve linear scalablity. |
|
N********n
发帖数: 8363 | 35
In your own linux propaganda article from lwn.net/Articles/411064/
it claims to run "thousands machines" to achieve "up to a million
tranctions per second".
I'm gonna give you a break here, let's say it's just 2000 machines.
1M / 2000 = 500 transactions per second. The 8 sql server nodes OTOH
achieve 5000 / second, and that's 5000 / 8 = 600+ each.
So answer me please which is greater here: 600+ or 500? Are you sure
you have enough IQ to decide? How about a calculator? LOL
You are hilarious. MSF... 阅读全帖 |
|
g*****g
发帖数: 34805 | 36
I bashed JS the language. But I've been calling JS to stay in web
programming all the time. It's all about practical. There're hundreds of
posts
in history to check. You shameless liar.
It's not a spin, LSE dumped windows because it never met their latency
requirement. And it cannot be fixed. Linux did, on the first day. Crash
as bad as it is, is a fixable issue.
You call NASDAQ stock exchange using SQL server. and no, that's not
stock exchange, that's a small system handling 5K transaction/s a... 阅读全帖 |
|
g*****g
发帖数: 34805 | 37 您老不知所云，我提的IOC, AOP, ORM，你找出来了吗？Google Search是一个计算复杂
而非逻辑复杂的应用，而且起步的时候Java还什么都不是。论Scalablity，现有的应用
里，Twitter大约是峰值负荷最高的。用户数多而且有关联，支持实时搜索。不像
Google search的内容不需要实时，数据可以复制，用户之间相互不影响。Ebay和
Amazon是用户数多而且需要
支持transaction的大型系统，相似级别的应用，不仅C++没有，C#也没有。我提的这些
都是Java或者JVM语言为主的混合应用。
所有的大型系统，最关键的在于数据库的处理。如果你要跟我讨论Scalability，我建
议你
先学习一下什么叫做CAP theorem。你举的例子就说明你在这方面完全没有经验。
么？ |
|
h*******n
发帖数: 82 | 38 你说再多JAVA FRAMEWORK也好，scale ability名词也好，什么30M用户同时收到消息也
好，你所做的就是个低级的工作，管理好你的用户连接池，发送PLAIN TEXT， 我也不
想和你辩论网游里面数据传送，容错机制之类的了。就是想让你这种做网站的自以为是
的人知道，你再吹捧JAVA多好，你的网站有多少用户，你堆一大堆垃圾设计模式也改变
不了你们做的是发发文字，做做加减乘除的低等事情的事实。还有scalability不是你
做屁网站专业的专有名词，别人用Multi-GPU scalability 破解密码，渲染图像，物理
模拟什么不比你发text强, 我一两块小显卡同时能handle的Tasks比你几十台服务器都
强。
latency
handle
users.
users |
|
g*****g
发帖数: 34805 | 39 你不懂装懂有意义吗？你要鼓吹Multi-GPU Scalability，难道MMO是合适的例子？
啥时候打MMO还需要图形工作站不成？好吧，就算MMO需要图形工作站，难道你所说的
几M在线用户在服务器端是用GPU处理的？
我老好心给你讲performance跟Scalablity的区别，听不听的进去是你的事情。
你满口垃圾，只能显得你弱。一开口就露怯，我老何必跟你一般见识。 |
|
g*****g
发帖数: 34805 | 40 您老不知所云，我提的IOC, AOP, ORM，你找出来了吗？Google Search是一个计算复杂
而非逻辑复杂的应用，而且起步的时候Java还什么都不是。论Scalablity，现有的应用
里，Twitter大约是峰值负荷最高的。用户数多而且有关联，支持实时搜索。不像
Google search的内容不需要实时，数据可以复制，用户之间相互不影响。Ebay和
Amazon是用户数多而且需要
支持transaction的大型系统，相似级别的应用，不仅C++没有，C#也没有。我提的这些
都是Java或者JVM语言为主的混合应用。
所有的大型系统，最关键的在于数据库的处理。如果你要跟我讨论Scalability，我建
议你
先学习一下什么叫做CAP theorem。你举的例子就说明你在这方面完全没有经验。
么？ |
|
h*******n
发帖数: 82 | 41 你说再多JAVA FRAMEWORK也好，scale ability名词也好，什么30M用户同时收到消息也
好，你所做的就是个低级的工作，管理好你的用户连接池，发送PLAIN TEXT， 我也不
想和你辩论网游里面数据传送，容错机制之类的了。就是想让你这种做网站的自以为是
的人知道，你再吹捧JAVA多好，你的网站有多少用户，你堆一大堆垃圾设计模式也改变
不了你们做的是发发文字，做做加减乘除的低等事情的事实。还有scalability不是你
做屁网站专业的专有名词，别人用Multi-GPU scalability 破解密码，渲染图像，物理
模拟什么不比你发text强, 我一两块小显卡同时能handle的Tasks比你几十台服务器都
强。
latency
handle
users.
users |
|
g*****g
发帖数: 34805 | 42 你不懂装懂有意义吗？你要鼓吹Multi-GPU Scalability，难道MMO是合适的例子？
啥时候打MMO还需要图形工作站不成？好吧，就算MMO需要图形工作站，难道你所说的
几M在线用户在服务器端是用GPU处理的？
我老好心给你讲performance跟Scalablity的区别，听不听的进去是你的事情。
你满口垃圾，只能显得你弱。一开口就露怯，我老何必跟你一般见识。 |
|
h******u
发帖数: 155 | 43 "大部分JAVA的对象分配通常都是几条机器指令"有點誇張，呵呵。目前最efficient也
就是thread-local bump-the-pointer allocation, 就算是fast path也需要平均 12
instructions。C/C++的memory allocation也不是一定需要freelist-based，很多
customized的allocator就用bump-the-pointer。並且allocation的efficiency不光光
是通過allocation幾條指令來衡量的，還有好多好多的因素，比如説cache locality,
同一個processor的不同threads會不會經常access同一個cache line，還有
fragmentation等等。各種原因在一起非常複雜。
另外performance-wise“可以肆无忌惮地创建对象，把一切对象化”不是一個good
idea :) 現在看到很多server applications的scalability已經被excessive object
creation所限制。我們study過... 阅读全帖 |
|
h******u
发帖数: 155 | 44 要看什么样的 application？不是什么程序都好fix的。并且，我前面主要在讲用java
实现的
data-processing system,例如hadoop,Giraph,Hyracks之类,不是简单query db的程序
。这些大程序往往需要在5G的memory上面处理500G的data stream，如果你看到所有的
interface都需要object, 这个程序很难去大大的优化。object pooling是一个很ad-
hoc的pattern，并不能fundamentally解决object带来的 overhead。一个很关键的做法
（现在很多系统都在用），就是不让系统里面object的数量随着数据量线形增长，否则
不可能scalable。怎么办，Hyracks用了这种做法：用buffer-based data management
。创建大的buffer，把数据放在buffer里面，而不要create小objects。 类似于region
-based memory management。 只有data processor （比如hash function, so... 阅读全帖 |
|
