x*********n
发帖数: 28013 | 1 在论坛和人争辩。
有点不太理解。
他说CSPF还要配合RSVP+IGP才能完成MPLS-TE
我觉得CSPF本来就建立LSP了，+IGP就直接可以MPLS-TE了。
不是太理解，大牛们，给解释一下好吗？ |
|
|
b****m
发帖数: 318 | 3 in BIOS setting, I can see the option for the iGP shared memory is only 128M
at max, however the MoBo claimed can be up to 1.7G.
I really can't find other options in BIOS, and my WEI for Aero is only 1.9..
.....
please help.
BTW, if I set the iGP to support multi-monitor, the shared memory can be set
to 64m only........ |
|
x*********n
发帖数: 28013 | 4 小弟现在看BGP看的有点苦恼，
有些问题detail不是很清楚，看到一半总会出来点redistribution啦，这个routing
table啦，那个BGP table啦，越弄越糊涂了。
是不是要先看一遍IGP，至少把ospf搞明白了，在看BGP会效率高一点？ |
|
s******v
发帖数: 4495 | 5 bgp tablet太大，ipv4 unicast有300-400K，会撑死IGP的 |
|
s******v
发帖数: 4495 | 6 有multicast, ipv6, vpnv4 unicast, l2vpn vpls, 这些都是SP上面的常有的
SP里面的IGP就是carry PE/ASBR loopback 这些NH reachability
不过你这么问，应该是enterprise里面的用法， |
|
x*********n
发帖数: 28013 | 7 哦。。我在理解理解。 脑子有点混乱。
SP 3种vpn， layer 2vpn，layer 3 vpn， vpls，
常用的mpls vpn算是layer2吗？，底层跑IGP，take care ipv4，上面跑mp-BGP，take
care vpnv4，
你说的multicast是说非vpn应用下的吗？multicast不是有其他的protocol吗？BGP
multicast干嘛啊？
有点不太理解。。 |
|
s*****g
发帖数: 1055 | 8 Let me try to answer your questions, there are two types of MPLS VPNs, MPLS
L2VPN and MPLS L3VPN, MPLS L3VPN is standardized in well known RFC2547,
payload is IP, MPLS L2VPN is a general term for (Cisco)AToM, EoMPLS and
VPLS, payload is FR/HDLC but more commonly Ethernet/Ethernet VLAN.
ISPs need to run IGP in their own internal network, because MPLS LSPs are
typically among iBGP loopbacks, iBGP has to be on top of TCP/IP, of course
you can have MPLS LSPs that are purely used to transport IP tra... 阅读全帖 |
|
s*****g
发帖数: 1055 | 9 You need to get basic concepts right first: 1)IGP LSAs carry bandwidth
information of interfaces, 2)CSPF calculates constrained path to destination
,3)RSVP signals the LSP along the path calculated in 2nd step.
You don't need to be a big-cow to get the basic concepts right, say by
spending one hour googling. |
|
x*********n
发帖数: 28013 | 10 哦。。明白了。。。
我还以为IGP先跑好，然后CSPF算好，直接signal好，然后transport label就可以哗哗
哗的跑了。
还要RSVP，好麻烦啊。
destination |
|
|
s***s
发帖数: 1301 | 12 刚才看到这个，说不定能解释双显卡问题
http://superuser.com/questions/282734/how-to-disable-nvidia-opt
Short answer: It doesn't "switch" and you can't disable the Intel IGP.
Long answer: From a hardware point of view the Intel IGP is always handling
image output to the laptop's LCD screen. The NVIDIA GPU copies rendered
graphics into the Intel IGP's frame buffer (which resides in system memory)
when it is active. The Optimus software makes certain programs use the GPU
based on the loaded profiles and user settings.... 阅读全帖 |
|
s**********9
发帖数: 1238 | 13
呵呵，我的意思是每个AS内的IGP路由目前ISP基本都是ISIS了，有3万条就足够了。
我想LZ可能看错了，500K条？那还要BGP来干吗？全用IGP就可以了嘛。目前你都查到
BGP路由有300K左右了，全换OSPF不好么？
就是因为IGP消耗大，支持不了那么多的路由，才有BGP的出现
链路状态协议的OSPF对CPU的开销比BGP要大多了。
实际上，50K的OSPF应该也没有的，我记的OSPF最多支持15000条路由
具体查查 |
|
s*****g
发帖数: 1055 | 14 There is no difference operation wise between ordered control vs.
independent control, just one extra configuration command, both modes are
equivalent. LDP has no idea and does not care what mode its neighbor is
running. Having said that, there is really not much practical use to have
any non/32 FECs. Only thing BGP/IGP shortcuts, VPN transport LSP need is a /
32 FEC
IGP/LDP synchronization has nothing to do with ordered or
independent control or size of LFIB. I believe IOS by default enables IG... 阅读全帖 |
|
n**********l
发帖数: 271 | 15 origin is set by the origin AS, usually meaningless for traffic engineering.
Different vendors use origin differently.
A common practice is to rewrite origin in all received routes.
core1.fmt1.he.net> show ip bgp routes detail 8.8.8.8
Number of BGP Routes matching display condition : 15
S:SUPPRESSED F:FILTERED s:STALE
1 Prefix: 8.8.8.0/24, Status: BI, Age: 9d1h58m39s
NEXT_HOP: 184.105.224.254, Metric: 15, Learned from Peer: 216.218.
252.165 (6939)
LOCAL_PREF: 10... 阅读全帖 |
|
d******a
发帖数: 8 | 16 Title: Director, Engineering
Location: Taipei
Team Size: 130+
PREFERRED EXPERIENCE :
- BSEE/ MSEE
- 15+ years direct PC industry working experience
- CPU and Chipset design/support experience
- Motherboard/platform experience is a plus
- Good English communication and written skill
- 10+ years of team management experience
DESCRIPTION OF DUTIES IN ADDITION TO THOSE IN JOB DESCRIPTION :
- Provide leadership for platform engineering team in Taiwan, which covers
DT IGP, Mobile IGP, server CPU and |
|
u*v
发帖数: 18 | 17 IBGP can not resend the BGP table from one IBGP neighor to other IBGP neighbor
, so you must be IBGP full mesh or using BGP reflector. You need not
physically connect these IBGP neighbor, BGP is using TCP, so just let the IBGP
nei can reach each other is enough, normally, you can run IGP for it. One
more thing, BGP will not install BGP into routing table and announce to other
AS, only if it syn with IGP or
you run IBGP on these nodes of the path to destination. |
|
s*****g
发帖数: 1055 | 18 Of course you don't run IGP between to SPs for option C, this is a basic
requriement of any inter-AS, no IGP between ASBRs.
Problem with option B is also obvious, ASBRs have to be VPN aware (you need
to configure vrf/vpn address family under ASBR's bgp in order to exchange
VPN labels with neighbors) and ASBRs have to hold all VPN routes (just
imagine if you have a dozen customers and each customer wants to have whole
Internet routes ..., not many vendors on the market can do over 1M routes
hardw |
|
s*****g
发帖数: 1055 | 19 No, I was not talking about MPLS VPN label, I am talking about pure MPLS ...
The router only needs to assign labels for its IGP routes, it does not need separate labels for its BGP routes, it only needs to advertise ONE label for its BGP router ID.
Any router that has Internet routes installed in its IGP table, that network engineer should be fired.
especialy |
|
z**r
发帖数: 17771 | 20 the BGP routes will be installed into IP table eventually right? anyway....
let me clarify the question from different point of view then, say the whole
network has IGP only, all 300k prefixes are IGP prefixes, and it's
approaching to 1M now
..
need separate labels for its BGP routes, it only needs to advertise ONE
label for its BGP router ID. |
|
c*****i
发帖数: 631 | 21 if we have 300k igp routes say ospf routes, we need worry about ospf first...
I'm not sure about other edge devices, but for cisco, I don't think GSR/ASR
can support such large number of igp routes.
whole |
|
s*****g
发帖数: 1055 | 22 Technically I don't think there is a problem, but what for? we already have
LDP/RSVP, which rely on IGP to distribute transport labels, IBGP can not be
used as IGP. |
|
c*****i
发帖数: 631 | 23 pe-pe的rip是igp。 ibgp是mp-ibgp,传mpls vpn route/ lable用的。当然如果你有
full mesh的ipv ibgp，就不需要rip的。
话说mpls support rip as igp吗。
上次还被人问了个TE需不需要mpls,汗阿。 |
|
s*****g
发帖数: 1055 | 24 Come on guys ... you always need IGP full-mesh iBGP or not, think about it,
iBGP relies on TCP, without IGP, how do you route BGP's TCP taffic?
"TE需不需要mpls", what kind of question is that? |
|
s*******8
发帖数: 12734 | 25 最近发现我太菜了，IGP没学好。
前阵子买了一个Barns nobel的coupon，20刀，不知道买啥书好，IGP书有啥推荐的么？ |
|
z**r
发帖数: 17771 | 26 来自主题: EmergingNetworking版 - 一个面试题 一个网络要从IP core升级到MPLS core+L3VPN，这个core大约1000台PE，IGP大约30K。
有什么办法可以seamlessly把所有现有的CE router都升级到相应的VRF里？要求是在升
级的过程中，不能中断服务，当然，短时间的中断是可以忍受的，比如改变一个配置，
这个新配置生效的过程中，服务中断可以接受。
今天面试别人，team里另外一个家伙问的，俺似乎也想不起啥有效的措施，route
leaking between GRT and VRF似乎不太灵，主要涉及大量的IGP prefixes，只能用
routing protocol来leak，但这样的话，貌似只有加一个local的loop一端是VRF，一端
是GRT才可能行得通。
还有啥办法？ |
|
s*****g
发帖数: 1055 | 27 来自主题: EmergingNetworking版 - 一个面试题 Internet access inside VRFs? how scalable is that going to be? say customers want to have full Internet BGP feed, so each VRF will have 350K FIB entries, in that case how many VRFs can a typical PE provide? 3?
Also how is SP's IGP prefixes size relevant? SP's IGP prefixes can well be in RFC1918 space. If PE-CE connection is Ethernet, can't we configure VLAN/sub-interfaces on both sides, one VLAN/subinterface for L3 VPN access and the other one for Internet access? that way each VRF FIB can be sm... 阅读全帖 |
|
l***y
发帖数: 791 | 28 If in core network running a mix of the two mode (i.e. multi-vendor
environment) would it be any worse than running just just one? Junos only
supports ordered control but i think cisco runs indepedent.
my thinking is that at such time when new link comes up, you'd drop packets
for a few seconds regardless of which mode, just the location of the
blackholing would be different.
it seems that igp sync is good for these occassions, but I am guessing it
must require ordered control, and also depends ... 阅读全帖 |
|
s*****g
发帖数: 1055 | 29 Right LIB!=LFIB, but one is the other one's (exact) copy and they are always
in sync. In theory, you can not predict how many LIB entries you will ever
get from your downstreams, and you don't know how much process power your
neighbor has, IGP/LDP synchronization is designed to coordinate between to
LDP peers and IGP to avoid MPLS forwarding blackholing regardless label-
mapping size or control plane processing power. Does my reasoning make
logical sense?
When you say "I think", you'd better giv... 阅读全帖 |
|
s******v
发帖数: 4495 | 30 but ipv4 connectivity with label is carried by IGP.
when I did show ip bgp
PE1#sh ip bgp 44.100.100.2
BGP routing table entry for 44.100.100.2/32, version 5
Paths: (1 available, best #1, table default, RIB-failure(17))
Not advertised to any peer
Refresh Epoch 1
Local
44.100.100.2 (metric 2) from 44.100.100.2 (44.100.100.22)
Origin incomplete, metric 0, localpref 100, valid, internal, best
rx pathid: 0, tx pathid: 0x0
this ipv4 prefix is even marked as RIB f... 阅读全帖 |
|
s*****g
发帖数: 1055 | 31 The first scenario is BGP shortcut, the second one is ospf shortcut, totally
irrelevant to each other. Yes, A will take physical links unless TE tunnel
is
announced to IGP, and this TE tunnel won' t be usable/visible by A's
upstreams unless forwarding adjencency is configured in which case bi-
directional TE is needed to keep IGP link-state database consistent.
/html/mpls-config9.html |
|
x*********n
发帖数: 28013 | 32 这个面试等了1年多了，石沉大海不知为啥就找我了。感谢chee的题目，题目有些相似。
我把全过程写一下，面的优点压力山大啊。
介绍你的daily duty，design什么，那个project最复杂，最难，遇到什么问题。
技术开始
1.single mode vs multimode
我说length，价格
MM的length value是啥？我说400，但是不同vendor不一样的，他说他们looking for
300 meters，你说的也算对。不同vendor不一样的。
2.Circuit CRC value, how do they work
瞎扯了一下，我说看看CPU，看看有没有什么broadcast storm，sh in transceiver看
看信号lost多少
3.circuit is not up, what will you check,
瞎说，真心不太懂。扯了一点transitmitter，smartjack，interface。
4.HSRP vs GLBP
又准备
5.OSPF LSA types,
又准备
6.how LSA 5 to reach LSA... 阅读全帖 |
|
x*********n
发帖数: 28013 | 33 eBGP needs to directly connected, iBGP doesn't
eBGP 不directly connected也能peer，3种方式，一种是常见的emulti-hop，一种是
disconnected，一种是TTL-max啥的，总之都是改TTL。
BGP是over TCP的，你想啊，这都是在layer 4了，底层是要有IGP running的。
synchronation都是要和IGP match才能做BGP的。
你可以把interview题目发上来，大家帮你看看，弄清楚了，下次就不会错了。 |
|
r**********e
发帖数: 2821 | 34 65nm的45W Ahtlon+780G不是有人侧过低于30w吗？
45nm Athlon xII功耗只能更低。
anandtech系统的侧过IGP系统的功耗：
E7200+G35〉athlon+780G〉athlon+GF8200。
有网站搞个790fx比P45平台，没意义。要比也是x48比790fx，780G比G35orG45. X58+i7
功耗还
暴高呢。 所以一般IGP平台比比低功耗还有点意义。 |
|
a*********p
发帖数: 717 | 35 有3块显卡
8800GTS 512
9800GT
GTS 250
用在Asus P5N7A-VM（GeForce 9300/nForce 730i，禁用IGP)上，用9800GT和GTS250,3D
游戏和3D mark中途总是会黑屏或自动重启。但是用8800GTS就从来没有问题。驱动，系
统和其他硬件都一样。单独用IGP也没有问题。
8800GTS是最耗电和最热的，温度最高要到84度左右，而9800GT和GTS 250黑屏之前的温
度70度也不到。
本来怀疑两块显卡有问题，结果放到另外两台机器上也都没有问题（DFI AMD 790GX和
ASUS Intel P55)
怎么NV 的显卡在自家的芯片组有问题，在AMD和intel的芯片组上都好好的。
我觉得好像走到了死胡同，不知道是什么出了问题。望大家帮忙分析分析 |
|
r**********e
发帖数: 2821 | 36 看用途，intel IGP板子都是垃圾不也卖了这么多年，当办公和灌水机够了。现在要么
i7,i5,要么785G IGP平台，775都是鸡肋了。 |
|
f********2
发帖数: 368 | 37 我想攒台新机，做视频打游戏用，一定会买discrete graphics card
现在microcenter有比较好的sandy bridge的deal,但其自带的graphics芯片貌似很多余
。据说想要用上这个IGP,一定要买h67的主板并且把一个显示器接到主板上。但为什么
要这么做呢?
有没有办法同时用IGP和外接显卡，并且提高图形性能呢？ |
|
a***e
发帖数: 27968 | 38 P67允许改倍频
H67不许改倍频，频率没法锁
但是改频率影响几乎所有系统部件，稳定性更差，
还有igp的超频能力不如core
Z68的麻烦，估计是需要igp core独立改倍频
不过P还是支持解码单元的，算有点良心 |
|
|
m*****e
发帖数: 21 | 40 Mother Board: ABIT NF7-M
Chipset: NVDIA nForce2 IGP+MCP2, 8*AGP
Audio: 6-Channel AC 97 CODEC
10/100M LAN
USB 2.0
IEEE 1394
Video Card: Best Data GeForce4 400MX 8*AGP 64M
Hard Disk: Maxtor 160G 7200RPM 8M Buffer
CD RW/DVD ROM Combo: IDE4816C0
Memory: Cosair DDR 256M 3200C2PT 400MHz 32*64 184pin
Monitor: Dell 17" LCD
Processor: AMD 2.0G
7年前攒的，还能卖几个钱？ |
|
w*l
发帖数: 2550 | 41 12/15/2006
关于科学研究诚信的公开信 (傅新元等120人连署)
尊敬的：教育部部长周济 科技部部长徐冠华 中国科学院院长路甬祥 国家自然科学基
金委员会主任陈宜瑜
作为科技工作者，我们十分关切国内近来多起有关学术研究不端的指控事件。我们愿意
针对科学研究诚信的重要问题表示如下观点：
1. 我们反对任何形式的学术不端行为，并强烈主张一个科技工作者应当保持高标准的
学术道德规范。我们支持揭露任何形式的学术不端。与此同时，要确证学术不端的 案
例，我们必须遵循合法和适当的调查程序。例如，美国公共卫生署关于学术研究不端行
为的处理指南规定，对学术研究不端行为指控的调查，起初应当是非公开 的，其首先
应当由所涉及机构的内部委员会调查并处理。“无罪假定 (Innocent until proven
guilty)”是我们在调查学术不端的指控时必须遵守的一个重要原则。
2. 如果争论或指控不能够在该研究机构内部解决，那么，更高的主管部门和基金支持
机构或专业协会应当介入，建立外部调查委员会，直到问题得到圆满解决。在调查 期
间，不论是指控者或被指控者，都应受到合法保护，其身份应该保密。被指... 阅读全帖 |
|
c*********0
发帖数: 1676 | 42 LMAO - Laughing My As* Off
LOL - Laughing Out Loud
AFAIK =» As Far As I Know
AFK =» Away From Keyboard
ASAP =» As Soon As Possible
BAS =» Big A$$ Smile
BBL =» Be Back Later
BBN =» Bye Bye Now
BBS =» Be Back Soon
BEG =» Big Evil Grin
BF =» Boyfriend
BIBO =» Beer In, Beer Out
BRB =» Be Right Back
BTW =» By The Way
BWL =» Bursting With Laughter
C&G =» Chuckle and Grin
CICO =» Coffee In, Coffee Out
CID =» Crying In Disgr... 阅读全帖 |
|
m*******t
发帖数: 140 | 43 “STP主要就是避免Layer 2 Loop，有点link status的意味。这个Tree的概念跟ISIS或
OSPF很相像”
Wrong statement... Both are working on the path building, but nothing else
shares the same. The fundamental difference is STP allows broadcast loop
first, then build up the tree path by blocking ports, but loop shall never
happen under IGPs. That's the reason Cisco is proposing TRILL (ISIS based)
to replace the STP.
Availability |
|
|
l**n
发帖数: 7272 | 45 不知道中文技术名词怎么说。贻笑大方了。
For many video cards (AGP, PCIe), the video memory is directly accessible by
GPU only. In this sense, the video memory is not in the CPU's memory
address space.
For low end system with on-board integrated video (IGP), the story is
different. |
|
a***e
发帖数: 27968 | 46 there are more IGP out there.
pad is a lowend pc system :)
the back end display memory needed is about 8MB~16MB even for the high res,
could be easily build on chip or in package.
it is the memory used for the graphic front end computing a problem.
it might support a mode of low res and auto fill in pixels in such case.
by |
|
z**r
发帖数: 17771 | 47 I actually wanted to mean the IGP routes are exchanged between AS's in
option C, because an ASBR must have the knowledge of all loopbacks of the PE
routers within the AS, then uses eBGP to advertise them to other AS's. This
requires greater trust between 2 SP's.
Also a lot of situations that ppl use RR to improve the scalability, thus
the eBGP connections exist only between the blue RR in the blue AS and the
green RR in the green AS.
In option B, ASBR VPN awareness shouldn't be a problem at all, |
|
p***a
发帖数: 28 | 48 i agree with you more on this.
but actually different ASes can be in a singile IGP domain, of course, in
this case the ASes should belong to a same provider.
need
whole
PEs |
|
s*****g
发帖数: 1055 | 49 For IP FRR, this is true because IGP will converge no matter what, but for
MPLS facility-backup FRR, since the path is pre-determined in the ingress,
ingress may not implement re-optimization of protected LSP, i.e, protected
LSP may or may not be adaptive, so traffic may stay in bypass path,
bandwidth requirement is not guaranteed especially N LSPs are riding on the
same bypass.
For detour FRR, bandwidth constrain is not a problem by nature. |
|
