L******t
发帖数: 1985 | 1 Juniper have laid off ~100 or so employees mostly from their QFabric
engineering and product marketing teams.
Just weeks ago I saw Jnpr QFabric team hiring ads on Linkedin.com.. |
|
f*****m
发帖数: 416 | 2 TRILL/SPB/FabricPath 可以用在flat的或者hierarchy 的网络上. Qfabric是flat的一
种实现方式 |
|
s*****g
发帖数: 1055 | 3 Hmm, Juniper is giving up QFabric? not a good news for the company. |
|
l***y
发帖数: 791 | 4 请问版上有没有在tri-state,熟DC/QFabric的顾问啊。 |
|
s**********n
发帖数: 2343 | 5 索性再给你比一比 Juniper 吹得最凶的scaling 吧， 这里Qfabric 和 VDX 也是半斤
八两， 谁好谁坏
还很难说。 这个题目技术性有点强， 通俗地说， Qfabric 没有遵照标准，自己也不
明说是什么技术
细节， 但是效果好。 Vdx 遵照IEEE标准， 性能也还可以，不像juniper 认为得那样
什么都做不了。
具体可以看下面的讨论， ADAM = Juniper, BROOK = brocade.
http://blogs.forrester.com/andre_kindness/11-02-23-
junipers_qfabric_the_dark_horse_in_the_datacenter_fabric_race
ADAM->Well let's see, VDX can do very few switches in a "fabric" (I'm
thinking maybe 4,
correct me if i'm wrong)
BROOK-> Yeap, not close. On the first release we support... 阅读全帖 |
|
w*f
发帖数: 111 | 6 update
answer to "What is Juniper QFabric in layman's term?"
Rakesh Singh:
Qfabic is a new network architecture from Juniper which makes the whole data
center behave as if it were one giant switch.
There are 3 componets to QFabric: Node, Interconnect and Director. To take
the analogy of a typical Chassis-Switch, the QF/Nodes are the line card in
chassis, The QF/Interconnect is the Switch Fabric and the QF/Director is the
CPU card.
The typical chassis switch is limited to 8 or 16 line cards ... 阅读全帖 |
|
d****t
发帖数: 474 | 7 在很多方面jnpr qfabric秒杀brcd vdx
scaling, power, network complex, latency, layer 3 support...
半年之内，对brcd比较重要，现在都在建设data center，他们的产品是不是可以站住
脚，data
center network，有jnpr's qfabric, csco's ?fabric, brcd's etherfabric，还
有hp，如果能够打响的话，业务就可以重回增长。
他的fiber channel switch是老大，但是没有增长。
消息，发了
follow up的邮件也没有回复。请问在那里工作的大侠，或者了解他们流程的，我这种
情形是不是就
没戏啦？ |
|
d****t
发帖数: 474 | 8 QFabric <900ns any case.
VDX latency 600ns只能在group内，如果要crossing fabric，latency就不行了，这些
都是
architecture上的缺陷造成的，以后不改体系结构要提高几乎不可能。
等着看市场怎么反应就行了。
OK, 48 port, 365W vs. 60 port 360 W.
再说latency , Qfabric 强一些，
QFX3500 900nS,
VDX6720-24 , 600 ns.
VDX6720-60 , 1800nS for any ports, 600 ns within 10-port group.
这就是所谓秒杀？ |
|
z**r
发帖数: 17771 | 9 上个星期出差，和负责Qfabric的一个manager吃了个饭，不过没聊细节。大体上似乎是
把若干个switch用QFabric连起来，达到一个virtual switch的效果？而且没有地理位
置的限制？ |
|
L******t
发帖数: 1985 | 10 QFabric is still not the same thing as TRILL.
TRILL is two-tier; QFabric is one-tier, totally flat. But TRILL is not the
end of story on Cisco side. |
|
t*********e
发帖数: 1136 | 11 There might be the following issues with this model:
1. Wiring will be too complex. QFabric essentially externalizes the guts of
a switch. It is like telling the network admins to design and assemble a
switch themselves. Is that an attractive proposition? And I don't see what
advantage this has over normal high-density modular switches.
2. Latency and reliability will be a problem. Switch latency is approaching
zero. More wiring means slower switch. This is against the trend. The more
wiring als... 阅读全帖 |
|
j*a
发帖数: 14423 | 12 Nexus刚出来的时候也就一坨吧，谁敢用？
Q2/08出的 真正成熟被70% F100公司用上又是什么时候呢？
Juniper’s Project Stratus/QFabric for the data center: After 2.5 years and
several announcements, Juniper has not delivered on their promise that
QFabric is designed to scale “…to tens of thousands of 10GbE ports”.
Delivered...
Cisco announced Nexus in January 2008, delivered it on Q2 2008 and Cisco’s
Nexus data center switch family has been adopted by more than 70% of the
Fortune 100 |
|
s******v
发帖数: 4495 | 13 半年之内，对brcd比较重要，现在都在建设data center，他们的产品是不是可以站住
脚，data
center network，有jnpr's qfabric, csco's ?fabric, brcd's etherfabric，还
有hp，如果能够打响的话，业务就可以重回增长。
他的fiber channel switch是老大，但是没有增长。
消息，发了
follow up的邮件也没有回复。请问在那里工作的大侠，或者了解他们流程的，我这种
情形是不是就
没戏啦？ |
|
s**********n
发帖数: 2343 | 14 OK, 48 port, 365W vs. 60 port 360 W.
再说latency , Qfabric 强一些，
QFX3500 900nS,
VDX6720-24 , 600 ns.
VDX6720-60 , 1800nS for any ports, 600 ns within 10-port group.
这就是所谓秒杀？ |
|
|
|
a***n
发帖数: 262 | 17 My simple analogy
QFabric is like wireless controller based architecture+.
node is thin/thick AP - Aruba AP
interconnect is like controller - Aruba Controller
window/view like NMS - Aruba Airwave |
|
L******t
发帖数: 1985 | 18 Must be jnpr promoting QFabric based switch to zher's company.
Csco would be doing similar thing soon if not already I guess. |
|
a**********k
发帖数: 1953 | 19 All companies are working hard to provide that now.
For instances, JNPR's QFabric, BRCD/FDRY's VCS, CSCO's
nexus based solution etc. Any predication who will be
the winner in the Data center front? |
|
L******t
发帖数: 1985 | 20 If you have a sound confidence on the QFabric, you should be less optimistic
about Arista?
Of course there's an important assumption, that is bigger brothers will execute well. |
|
L******t
发帖数: 1985 | 21 What details do you know about Arista? Had some lengthy talks with M?
QFabric will at least match Arista on latency, right? Then what else
advantage will Arista have then? Price could be one I know..
Not dismissing Arista, just very curious. |
|
|
|
k*****s
发帖数: 231 | 24 Datacenter这一块，Virtual Fabric + LACP 应该会不错的。直接把网络变得很flat了
，然后主机来个vswitch.
大家维护3层路由以及很累很烦的了，再来个2层路由还不累坏我们？ |
|
z**r
发帖数: 17771 | 25 就是折腾，Layer3烦了，就上layer 2，过段时间问题出来了，然后再上layer 3，其实
看看现在的网络，就是这么绕圈子 |
|
a***n
发帖数: 262 | 26 Can not agree more :-)
centralize -> decentralize/distributed -> centralize
and spiral convolution
重在折腾
就是折腾，Layer3烦了，就上layer 2，过段时间问题出来了，然后再上layer 3，其实
看看现在的网络，就是这么绕圈子 |
|
A***i
发帖数: 17 | 27 Data Center switching的几大技术热点:
1.low latency： 基本来个cut-through L3 ASICs
2. IO consolidation： FCoE, iSCSI
3. Virtualization, Cisco vss/vpc, HP VC, Juniper Qfabric,
Brocade VCS,
Arista, vLAG, etc...目的是去掉STP， load balancing, full
redundancy, fast
convergence.
4. L2VPN (vmware, storage, etc need L2 spread over multi-
data center for
security and managment): MPLS(L2/L3 VPN, 已经做烂了），
Cisco OTV (mac-
based VPN)
这些技术基本都得构架在10G端口上实施.10G恐怕还得整上５－１０年，才能
让系统和
应用的开发，优化，整合,让大部分应用跟上１０G,否则vmware还会继续通
吃，把这些
个１０G分成１G... 阅读全帖 |
|
A***i
发帖数: 17 | 28 Data Center switching的几大技术热点:
1.low latency： 基本来个cut-through L3 ASICs
2. IO consolidation： FCoE, iSCSI
3. Virtualization, Cisco vss/vpc, HP VC, Juniper Qfabric,
Brocade VCS,
Arista, vLAG, etc...目的是去掉STP， load balancing, full
redundancy, fast
convergence.
4. L2VPN (vmware, storage, etc need L2 spread over multi-
data center for
security and managment): MPLS(L2/L3 VPN, 已经做烂了），
Cisco OTV (mac-
based VPN)
这些技术基本都得构架在10G端口上实施.10G恐怕还得整上５－１０年，才能
让系统和
应用的开发，优化，整合,让大部分应用跟上１０G,否则vmware还会继续通
吃，把这些
个１０G分成１G... 阅读全帖 |
|
|
s*****g
发帖数: 1055 | 30 I don't think openflow plays much role here, it is more about customer
acceptance and competition pressure.
Insiders please comment. |
|
|
c*a
发帖数: 806 | 32 如果是Fee based, 我不介意travel：） |
|
l***y
发帖数: 791 | 33 起码的几个月。他们还在确认这个budget.我是个菜鸟就是帮着他们问问。
有确切消息我再问各位高手吧！ |
|
|
c*a
发帖数: 806 | 35 晕，跟我抢生意:)
不过有人画了一张大饼，很诱人 |
|
|
|
r********t
发帖数: 338 | 38 Cisco 推fabric path/vPC+, Juniper 搞Qfabric, 加上VDC/Logical router，层次已
经很淡化了。但和客户聊大家还是习惯扯Core, access啥的，反正心知肚明。 |
|
r********t
发帖数: 338 | 39 来自主题: EmergingNetworking版 - 糗事都凑一块了 不会，那就咱一个老中。总算，毛病给重现了。BU咂咂呼呼，咱忍不住问测试的兄弟，
“FEX到底用了几个？”
下周上门吹SUP2e看样子很艰难哈。有点不想给C干了，可J也不怎样，Qfabric都快咽气
了。 |
|
r********t
发帖数: 338 | 40 SDN重写控制层，还是加一个薄层在现有的控制层上。比如，在SDN加条路由，是加入在
OSPF，还是自己在每个路由器上编程加入。前者，依旧依靠vendor，后者，风险很大，
Caspian Networks 和 J的Qfabric都是前车之鉴。 |
|
r********t
发帖数: 338 | 41 SDN重写控制层，还是加一个薄层在现有的控制层上。比如，在SDN加条路由，是加入在
OSPF，还是自己在每个路由器上编程加入。前者，依旧依靠vendor，后者，风险很大，
Caspian Networks 和 J的Qfabric都是前车之鉴。 |
|
a**********k
发帖数: 1953 | 42 According to those SDN pioneers, on top of the existing
control/data planes, there will be a NOS layer on centralized
controllers, and on top of that, there will be a network
virtualization layer, on which all future network software
applications will run.
Caspian was Larry Roberts's experiments on an immature idea,
and QFabric is more of a logic switch made up of three types
of separate, distributed components: control plane, line card
and backplane. Not really SDN IMO. |
|
r********t
发帖数: 338 | 43 vPC, VSS <=> virtual chassis, chassis cluster
FP/Trill <=> Qfabric |
|
|
L******t
发帖数: 1985 | 45 There's a come back. Juniper seems to be investing heavily on QFabric-II. |
|
b***p
发帖数: 700 | 46 最近看了几个training，难道DC用L3已经是大势？都不用L2，神马TRILL，QFabric全都
玩完了？
而且大家都用BRCM的Switch Chip，这个Vendor还有啥区别啊？
还有这个Trident和Jericho有啥子区别啊？是两个不同的系列，有没有这方面比较简单
的分析？ |
|
s*****g
发帖数: 1055 | 47 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。 |
|
b***p
发帖数: 700 | 48 最近看了几个training，难道DC用L3已经是大势？都不用L2，神马TRILL，QFabric全都
玩完了？
而且大家都用BRCM的Switch Chip，这个Vendor还有啥区别啊？
还有这个Trident和Jericho有啥子区别啊？是两个不同的系列，有没有这方面比较简单
的分析？ |
|
s*****g
发帖数: 1055 | 49 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. |
|
t******0
发帖数: 3757 | 50
L3 肯定是大势所趋。。EVPN直接到TOR！！！
不要浪费时间在神TRILL， QFabric。。。目前来看VxLAN终究也会随风而逝 |
|
