j********x
发帖数: 2330 | 1 网络拓扑设计 建模 渐进式部署 网络配置等等这些过程，有专门的流程方法么？ |
z**r
发帖数: 17771 | 2 多大网络，很小的就没必要走过程了
【在 j********x 的大作中提到】
: 网络拓扑设计 建模 渐进式部署 网络配置等等这些过程，有专门的流程方法么？
|
j********x
发帖数: 2330 | 3 很大吧，几万台server的规模，想知道市面上一般是怎么做的。 |
z**r
发帖数: 17771 | 4 如果是几万台物理server, 可能有几十万虚拟机，算中等规模。这个规模的网络设计起
来比较灵活，可以有很多选择。你是经手做这个项目还是就感兴趣随便一问？谁家的设
备？很难想象没有vendor帮忙设计
【在 j********x 的大作中提到】
: 很大吧，几万台server的规模，想知道市面上一般是怎么做的。
|
s*****g
发帖数: 1055 | 5 Also depend on what are the applications of the data center? raw
computing? cloud service? hosting? what are the requirements? LZ's question
is too broad.
I agree that LZ is asking the wrong question in the wrong board.
【在 z**r 的大作中提到】
: 如果是几万台物理server, 可能有几十万虚拟机，算中等规模。这个规模的网络设计起
: 来比较灵活，可以有很多选择。你是经手做这个项目还是就感兴趣随便一问？谁家的设
: 备？很难想象没有vendor帮忙设计
|
j********x
发帖数: 2330 | |
s*****g
发帖数: 1055 | 7 As Zher suggested, for this size of network, you should probably be issuing
RFPs to multiple vendors and ask them how they are going to design and
position their equipment, big players like Cisco will be able to give a
complete all Cisco design/solution, smaller players will usually partner
with other vendors to fill production gaps. Either approach has pros and
cons, depend on what is best for your network or your architects' favor.
You should also have in-house engineers who can talk intelligently to
vendors, so you don't get ripped off (for example, you are being ripped off
if Cisco sales position a CRS at your wiring closet), at least know each
product's strong and weak points which may or may not be relevant to your
network, be sure to ask vendor's road map so you don't get stuck on EoL soon
products.
【在 j********x 的大作中提到】
: 哪个board可以问？
|
