IP组播和基本组播（B组播）之间的区别

Question

What is the difference between multicast and Basic multicast(B-Multicast)?

How each of them works exactly?

---

This is related to distributed systems. We have basic multicast which just deliver the message with basic guarantees and reliability. Since the process is alive and doesn't crash the message will be delivered. However, if the process crashes there is no guarantee for reliability. Therefore, the term R-multicast(reliable multicast) was introduced. The R-multicast brings advance reliability by using the concepts of ATOMIC. It relies on some features such as integrity, Validity. Moreover, there is an agreement which says each of the members of the group also multicast the message after being delivered. This agreement brings 100% integrity.In addition, it can be implemented over either B-multicast or IP multicast such as R-IP multicast vs RB Multicast. In RB multicast the sender will be imploded by acks when the infra is scalable whereas in R multicast over IP multicast since the omissionn and failure of the message is low, due to the IP multicast tech, only NACK will be sent back. After all, I want to know the differences between IP multicast and B-multicast. How they exactly work? I know the concepts as is written but need more detail.

Answer 1

The difference between IP multicast and basic multicast is related to where in the network stack the multicast is occurring.

In the case of IP multicast, this occurs at the network layer, ie IP. For IPv4, addresses in the range 224.0.0.0/4 ( 224.0.0.0 - 239.255.255.255 ) are multicast addresses. For IPv6, the addresses are in the FFxx::/16 range. Packets with destination addresses in this range are sent to multiple hosts as requested via IGMP messages.

Basic multicast refers to multicasting at the link layer or below. In the case of Ethernet, any frame where the least significant bit of the first byte of the destination address are sent to all nodes in the network.

When Ethernet frames carry IPv4 multicast data, the destination addresses are in the range 01:00:5E:00:00:00 - 01:00:5E:7F:FF:FF , while frames with IPv6 multicast data have addresses in the 33:33:xx:xx:xx:xx range. There are other addresses used for other link-layer protocols, such as spanning tree or Ethernet flow control.

Some physical layers such as terrestrial wireless or satellite are inherently broadcast.

See here more detail on different types of multicast addresses .

As far as reliable multicast goes, that's an application level construct built on top of multicast. Since IP is inherently unreliable, reliable multicast built on top of IP needs to account for this unreliability. Most multicast technologies at lower levels are also unreliable and thus need application layer reliability as well.

Answer 2

As for IPv4 -- it's similar in IPv6

Multicast addresses Theses correspond to the class D. The class D:

• First octet: 224 - 239
• First octet pattern: 1110*
• These IP addresses are multicast addresses.

there are used for specific (routing protocol, service discovery, NTP), sometimes experimental, use-cases.

Network nodes must perform a join(multicast-address) call in order to receive packet sent to the address multicast-address . There can be many multicast addresses present in a network.

Broadcast address

There is only one broadcast address in every network. This address is constructed with all its host-part IP address bits set to 1 .

If the network is 192.168.0.0/24 the last octet is the host-part IP address (and the first three ones are the network-part IP address). The broadcast address is then 192.168.0. 255 .

The broadcast address is used to send packet to all the nodes within the LAN , not further, not only the nodes that would had performed a join(multicast-address) call -- that make no sense.

---

More details on this answer .