2017年６月に発表されたtrident 3、trident 2+の後継と言うよりは。Cavium XPliant等が特徴としているプログラマビリティとテレメトリー機能を特徴としているようです。
従来の流れだと、2017年年末にはこのチップを使用した製品が発表されるでしょう。Edgecore AS7726 and AS7326がTrident 3搭載 (2019/04/03追記)
|
trident 2+
BCM56860
|
trident 3
BCM56870
|
trident 3
BCM56873
|
|
|---|---|---|---|
| ファブリック容量 | 1.28Tbps | 3.2Tbps | 2.0T |
| SerDes | 128x 25G | ||
| port構成 |
40G x 32
10G x 48 X 40G x 6
|
100Gx32
25G x 48 + 100G x 8
|
10G x 48 + 100G x 6 |
| host CPU interface | PCIe Gen3 x4 | PCIe Gen3 x4 | |
| buffer | 12MB | 32MB | 32MB |
| IPv4 table | 64K | ||
| IPv6 table | 20K |
trident 3ファミリーには上記の他にも200Gと小容量の物も用意され、wifiのアクセスポイントなどをターゲットにしているようです。
cumulusのドキュメントによるとフィルターの数などはtrident2+と3は差が無いようです。(2019/04/03追記)
https://docs.cumulusnetworks.com/display/DOCS/Netfilter+-+ACLs
経路数も差が無し
https://docs.cumulusnetworks.com/display/DOCS/Routing
これからはプログラマブルなシリコンが一般的になるでしょう。
overlay / encap
PPPoE, QUIC,VXLAN-GPE,Geneve, ILA
VXLAN GPE(Generic Protocol Extension)
eternet以外をカプセル化するフォーマット。ダイレクトにIPv4,IPv6とか
NSH(Network Service Header)
programmable pipeline
five stage、ファブリックの前に三段、後に二段。
telemetry
https://youtu.be/Wp9wtS-e1QU
Tomahawk® 3
| Tomahawk® | Tomahawk® 2 | Tomahawk® 3 | ||
|---|---|---|---|---|
| BCM56960 | BCM56970 |
BCM56980
BCM56982
|
||
| capacity | 3.2T | 6.4T | 12.8T | |
| typical config | 100G x 32 | 100G x 64 | 400G x 32 | |
| host CPU interface | Gen2 x4 | PCIe Gen3 x4 | ||
| packet buffer | 85 MB | |||
| SERDES | 128 x 25Gbps | 256 x 25Gbps | 256 x 56G-PAM4 | |
| process | 28nm | 16nm | 16nm | 7nm |
| annaunce | 2014/09/24 | 2016/10/11 | 2017/12/19 | 2019? |
https://jp.broadcom.com/company/news/product-releases/2323373
cisco nexus 3100-Z
Edgecore AS9716-32X,
ARISTA 7060,
http://blog.ipspace.net/2015/12/broadcom-tomahawk-101.html
- Similar to Trident-2, Tomahawk becomes line-rate (3.2 Tbps) at packet sizes above 250 bytes;
- Presentation claims overlay routing (VXLAN-to-VXLAN or VXLAN-to-VLAN) is not supported, which is a bit surprising as the forwarding pipeline includes tunnel termination before L2 and L3 lookup, which should be good enough;
- The switching silicon has 10 queues per port (nice!);
- Switching latency is approximately 500 ns and can be reduced to 300 ns if the chipset is reconfigured into doing only simple L2 switching;
- Unified forwarding table (UFT; 128K entries) is split in memory banks that can be allocated to L2 entries, ARP entries and L3 LPM entries;
- One of the printouts in the presentation hinted at 1K LPM IPv6 prefixes longer than /64;
- Tomahawk support exact matching of ACL entries in UFT (not TCAM). UFT split with filter-mode profile can have 64K ACL entries, 16K IP LPM entries and 8K ARP/MAC entries;
- There are 43 queues between the switching silicon and CPU, and you can configure control-plane policing parameters on each queue;
- The hardware supports 16K MPLS labels (must be a separate MPLS lookup table, not TCAM tricks);
- TCAM slicing is too tricky for me to understand, but it seems you’ll get between 512 and 6K TCAM entries based on the complexity of the matching conditions. Based on the matching length used by Junos you get up to 512 port- or VLAN ACL entries or up to 1024 IP ACL entries;
- TCAM is not wide enough for all possible IPv6 matching conditions, so the hardware uses address compression. It seems you can have at most 128 source and destination IPv6 addresses in all filters deployed on the box;
Mellanox Spectrum-2
2017/07/07アナウンス
400G x 16 port
http://ir.mellanox.com/releasedetail.cfm?ReleaseID=1032283
CAVIUM XPliant
| capacity | SERDES | release | |||
|---|---|---|---|---|---|
| XP80 | CNX880XX | 3.2Tbps | 25G x n | 28nm | 2014/09 |
| XP60 | CNX680XX | 720Gbps | 10G x n | 28nm | 2017/07 |
| XP70 | CNX700XX | 1.4Tbps | 25G x n | 28nm | 2017/07 |
XP70はDELL SonicWALLに使われている？
XP80(CNX88091)はEdgeCore AS7512に搭載されている
XP80(CNX88091)はWedge 100Cに搭載されている
XP80はArista 7160に搭載されている