 |
 General |
1. |
Can you provide detailed information regarding the CPU, flash, SDRAM, packet memory, and backplane bandwidth of the LEO standalone ES3526X? |
Answer:
| CPU |
Built in 32 bit ARM 7 CPU (100 MHz ) |
| CPU cache |
8 KB I/D cache |
| Flash code space |
8MB |
| Processor DRAM size |
32MB |
| Internal packet memory |
512KB ( 4Mb ) |
| Extended packet memory |
No |
| Back plane bandwidth |
125MHz × 256 bit = 32Gbps |
|
| 2. |
What is the hardware architecture of the LEO Standalone ES3526X? |
Answer:
( Please click every part and see the title.)
|
| 3. |
What kinds of extension modules do the LEO standalone ES3526X support? |
Answer:
LEO standalone supports 1 port 1000SX, 1000LX, 1000T gigabit modules and 1 port 100FX(MMF),100FX(SMF) modules.
|
| 4. |
Are the extension modules hot swappable? |
Answer:
No, the optional modules do not support hot swap, when you remove or insert the module while the switch is powered on, the system will go down.
|
| 5. |
Does the embedded ARM CPU contain MMU (memory protection unit)? |
Answer:
No, The embedded ARM CPU is ARM7TDMI, which does not have an MMU.
The memory controller in Leo, called MPMC provides caching, DRAM, SRAM, Flash memory interface, but no virtual memory nor memory protection.
|
|
| Rate Limiting |
| 1. |
Can the LEO standalone support both ingress and egress rate limiting? If so, what is the granularity; and how many levels are supported? |
Answer:
LEO standalone support both Ingress/Egress rate limiting. There are 2 kinds of granularities for 10/100Mbps connection, 512Kbps and 3.3Mbps. There is only one kind of granularity 33.3Mbps for 1000Mbps connection.
Under each granularity, there are 30 levels rate limiting scales showing below:
Rate limiting scale N*512Kbps (N=1-30)
Rate limiting scale N*3.3Mbps (N=1-30)
Rate limiting scale N*3.3Mbps (N=1-30)
|
| 2. |
Can the LEO standalone ES3526X support 1Mbps granularity rate limiting? |
Answer:
The switch controller chipset does support 1Mbps granularity rate limiting, but the total level will remain at 30, which means that the highest bandwidth is 30Mbps. Even if the port with rate limiting is disabled, the highest bandwidth is still 30Mbps.
We suggest the user uses 3.3Mbps granularity, so that the highest bandwidth will be 100Mbps for both level 30(30x3.3Mbps) and the rate-limiting disabled port.
Although the specs are impressive, it would be very impractical for other vendors who claim to support 1Mbps granularity and 100 levels, because the ISP or carrier won't be able to classify customer levels to 100 scales and have 100 kinds of pay rates.
|
|
| Port Mirroring |
| 1. |
After configured port mirroring, the listening port loss its switching function, why? |
Answer:
This is chip's behavior to avoid local traffic interference while monitoring other port.
|
| 2. |
Can Port Mirroring monitor both TX & RX traffic simultaneously? |
Answer:
No, Port mirroring can only monitor TX or RX once at a time.
|
| 3. |
Can Port Mirroring monitor two TX/RX ports to one monitoring port? |
Answer:
No, Port mirroring can only monitor one port at a time.
|
|
| Access Control List |
| 1. |
What is the maximum number of filters (IP, extended and MAC filter)? |
Answer:
The maximum number of access control list is 32. Each access control list can contain 32 access control entries. The total number of access control entries can't exceed 94.
|
| 2. |
Why does the chipset support 128 Flow Classifier rules when the total number of access control entries is only 94? |
Answer:
Although the chipset supports 128 multi-field classification rules, one rule is used as the default rule that permits access without changing/checking anything. The other 32 rules were used for TCP port COS mapping.
128-1(default rule)-32(TCP port to COS) - 1(MAC Deny rule) = 94
|
| 3. |
When configuring the MAC ACL to block PC1, the traffic blocking only worked when the packet contained source MAC (PC1) & destination MAC (broadcast) codes. In other words PC1 cannot ping PC2 when PC1 is unable to see PC2 or the arp cache of PC1 age out, only if you add a static arp, can PC1 ping PC2. Why does this happen? |
Answer:
If you add a static ARP then PC1 and PC2 will use ICMP packet, in LEO this ICMP packet will be classified as an IP rule not a MAC rule so PC1 can still ping PC2 if static ARP is configured on PC1.
|
| 4, |
How can I configure the NetBIOS filter (LLC,IPX,UDP,TCP)? |
Answer:
MAC access-list mac_deny:
deny any any
permit any any ethertype 0806
permit any any ethertype 0800
IP extended access-list all_deny:
deny tcp any any destination-port 137
deny tcp any any destination-port 138
deny tcp any any destination-port 139
deny tcp any any destination-port 445
deny udp any any destination-port 137
deny udp any any destination-port 138
deny udp any any destination-port 139
deny udp any any destination-port 445
|
| 5. |
Why can't I apply ACL to all ports? |
Answer:
If there are 9 rules in an access list that bind to 10 interfaces there will be 9X10=90 entries in the rule table. Since the maximum number of entries should not exceed 94, an error message will be shown if you are trying to apply the 11th port.
|
|
| Private VLAN |
| 1. |
What is isolated vlan? |
Answer:
When the VLAN is isolated, the member ports in the VLAN are unable to talk to each other and are only able to communicate with the uplink port.
This example shows how to configure ports 1-6 to the isolated VLAN 2 where ports 1-5 are downlink ports and port 6 is an uplink port.
Console(config)#vlan database
Console(config-vlan)#private-vlan 2 isolated
Console(config-vlan)#ex
Console(config)#int e 1/1-5
Console(config-if)#switchport mode private-vlan host
Console(config-if)#switchport private-vlan isolated 2
Console(config-if)#int e 1/6
Console(config-if)#sw mode private-vlan promiscuous
Console(config-if)#sw private-vlan isolated 2
|
| 2. |
How many isolated VLANs can be supported? |
Answer:
ES3526X can support only one isolated VLAN.
|
|
IEEE 802.1X |
| 1. |
Can I remove all ports from VLAN 1? |
Answer:
Because VLAN 1 is the management VLAN by default, at least one port belongs to VLAN 1. If you configure the IP interface in another VLAN, then you are able to remove all ports from VLAN 1.
|
|
| Management VLAN |
| 1. |
Can I remove all ports from VLAN 1? |
Answer:
Because VLAN 1 is the management VLAN by default, at least one port belongs to VLAN 1. If you configure the IP interface in another VLAN, then you are able to remove all ports from VLAN 1.
|
|
| IGMP Snooping |
| 1. |
How many multicast groups can be supported while using IGMP snooping? |
Answer:
The LEO standalone Supports 64 multicast groups.
|
|
|
| Link Aggregation |
| 1. |
Can ports with different duplex modes trunk together? |
Answer:
No, each member port of a trunk should have the same duplex mode: for example if port 23 have no negotiation, speed-duplex mode 100Full, and ports 21,22 24 have negotiation, therefore only port 24 can be a member of the trunk.
|
2.
|
If I created a 4-port trunk (21-24) either static or LACP on the ES3526X which then connected to another switch that sent traffic with smart bits from port 1,2,3 on one switch to port 1,2,3 of the other switch with the following addresses.
Will the first 500 streams of traffic always go through port 22 and 24;
Will the second 500 streams of traffic always go through port 23;
Will the third 500 streams traffic always go through port 22 and 24;
and will no traffic go through port 21?
src 000000040001~0401f4(500EA)-->dst 000000010001~0101f4 (500EA) ---->port 22,24
src 000000050001~0501f4(500EA)-->dst 000000030001~0301f4(500EA) ---->port 23
src 000000060001~0601f4(500EA)-->dst 000000030001~0301f4(500EA) ---->port 22,24 |
Answer:
We implement both source & destination MAC as hash sources, because the source MAC and destination MAC both increase by 1 each time, after hashing by the CRC-12, the value will always remain the same, so they will always send traffic via same port of a trunk.
For proper distribution, try to increase the destination MAC only.
|
|
|
Technology Focus 
Whitepaper Download 
