Network Working Group Danny McPherson
INTERNET DRAFT Amber Networks, Inc.
Barry Dykes
December 2000 Onesecure, Inc.
VLAN Aggregation for Efficient IP Address Allocation
<draft-mcpherson-vlan-ipagg-01.txt>
1. Status of this Memo
This document is an Internet-Draft and is in full conformance with
all provisions of Section 10 of RFC 2026.
Internet-Drafts are working documents of the Internet Engineering
Task Force (IETF), its areas, and its working groups. Note that
other groups may also distribute working documents as Internet-
Drafts.
Internet-Drafts are draft documents valid for a maximum of six months
and may be updated, replaced, or obsoleted by other documents at any
time. It is inappropriate to use Internet- Drafts as reference
material or to cite them other than as "work in progress."
The list of current Internet-Drafts can be accessed at
http://www.ietf.org/ietf/1id-abstracts.txt
The list of Internet-Draft Shadow Directories can be accessed at
http://www.ietf.org/shadow.html.
2. Abstract
This document introduces the concept of VLAN [802.1Q] aggregation as
it relates to IPv4 address allocation. A mechanism is described by
which hosts that reside in the same physical switched infrastructure,
but separate virtual broadcast domains, are addressed from the same
IPv4 subnet and share a common default gateway IP address, thereby
removing the requirement of a dedicated IP subnet for each virtual
Local Area Network (LAN) or Metropolitan Area Network (MAN).
Employing such a mechanism significantly decreases IPv4 address
consumption in virtual LANs and MANs. It may also ease
administration overhead of IPv4 addresses within the network.
McPherson, Dykes [Page 1]
INTERNET DRAFT December 2000
3. Specification of Requirements
The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
"SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this
document are to be interpreted as described in [RFC 2119].
4. Introduction
The VLAN aggregation technique described in this document provides a
mechanism by which hosts that reside within the same physical
switched infrastructure, but separate virtual broadcast domains, MAY
be addressed from the same IPv4 subnet and MAY share a common default
gateway IPv4 address.
Such a mechanism provides several advantages over traditional IPv4
addressing architectures employed in large switched LANs today. The
primary advantage, that of IPv4 address space conservation, can be
realized when considering the diagram in Figure 1:
Figure 1:
+------+ +------+ +------+ +------+ +------+
| | | | | | | | | |
| A.1 | | A.2 | | B.1 | | C.1 | | B.2 |
| | | | | | | | | |
+------+ +------+ +------+ +------+ +------+
| | | /
| | | /
+-----------------------------------+ /
| |
| Ethernet Switch(es) |
| |
+-----------------------------------+
|
|
+--------+
| |
| Router |
| |
+--------+
In the Figure 1 hosts A.1 and A.2 belong to customer A, VLAN A.
Hosts B.1 and B.2 belong to customer B, VLAN B. Host C.1 belongs to
customer C and resides in it's own virtual LAN, VLAN C.
Traditionally, an IP subnet would be allocated for each customer,
based on initial IP requirements for address space utilization, as
McPherson, Dykes [Page 2]
INTERNET DRAFT December 2000
well as on projections of future utilization. For example, a scheme
such as that illustrated in Table 1 may be used.
Table 1:
Gateway Usable Customer
Customer IP Subnet Address Hosts Hosts
======== ============ ======= ====== ========
A 1.1.1.0/28 1.1.1.1 14 13
B 1.1.1.16/29 1.1.1.17 6 5
C 1.1.1.24/30 1.1.1.25 2 1
Customer A's initial deployment consists of 2 hosts, though they
project growth of up to 10 hosts. As a result, they're allocated
the IP subnet 1.1.1.0/28 which provides 16 IP addresses. The first
IP address, 1.1.1.0, represents the subnetwork number. The last IP
address, 1.1.1.15, represents the directed broadcast address. The
first usable address of the subnet, 1.1.1.1, is assigned to the
router and serves as the default gateway IP address for the subnet.
The customer is left 13 IP addresses, even though their requirement
was only for 10 IP addresses.
Customer B's initial deployment consists of 2 hosts, though they
project growth of up to 5 hosts. As a result, they're allocated the
IP subnet 1.1.1.16/29 which provides 8 IP addresses. The first IP
address, 1.1.1.16, represents the subnetwork number. The last IP
address, 1.1.1.23, represents the directed broadcast address. The
first usable address of the subnet, 1.1.1.17, is assigned to the
router and serves as the default gateway IP address for the subnet.
The customer is left 5 with IP addresses.
Customer C's initial deployment consists of 1 host, and they have no
plans of deploying additional hosts. As a result, they're allocated
the IP subnet 1.1.1.24/30 which provides 4 IP addresses. The first
IP address, 1.1.1.24, represents the subnetwork number. The last IP
address, 1.1.1.27, represents the directed broadcast address. The
first usable address of the subnet, 1.1.1.25, is assigned to the
router and serves as the default gateway IP address for the subnet.
The customer is left 1 IP address.
The sum of address requirements for all three customers is 16. The
most optimal address allocation scheme here requires 28 IP addresses.
Now, if customer A only grows to use 3 of his available address, the
additional IP addresses can't be used for other customers.
Also, assume customer C determines the need to deploy one additional
host, and as such, requires one additional IP address. Because all
of the addresses within the existing IP subnet 1.1.1.24/30 are used,
McPherson, Dykes [Page 3]
INTERNET DRAFT December 2000
and the following address space has been allocated to other
customers, a new subnet is required. Ideally, the customer would be
allocated a /29 and renumber host C.1 into the new subnet. However,
the customer is of the opinion that renumbering is not a viable
option. As such, another IP subnet is allocated to the customer,
this time perhaps a /29, providing two additional addresses for
future use.
As you can see, the number of IP addresses consumed by the subnetwork
number, directed broadcast address, and a unique gateway address for
each subnet is quite significant. Also, the inherent constraints of
the addressing architecture significantly reduce flexibility.
5. Discussion
If within the switched environment, on the routed side of the
network, we introduce the notion of sub-VLANs and super-VLANs, a much
more optimal approach to IP addressing can be realized.
Essentially, what occurs is that each sub-VLAN (customer) remains
within a separate broadcast domain. One or more sub-VLANs belong to
a super-VLAN, and utilize the default gateway IP address of the
super-VLAN. Hosts within the sub-VLANs are numbered out of IP
subnets associated with the super-VLAN, and their IP subnet masking
information reflects that of the super-VLAN subnet.
If desired, the super-VLAN router performs functions similar to Proxy
ARP to enable communication between hosts that are members of
different sub-VLANs.
This model results in a much more efficient address allocation
architecture. It also provides network operators with a mechanism to
provide standard default gateway address assignments.
Let's again consider Figure 1, now utilizing the super-VLAN sub-VLAN
model. Table 2 provides the new addressing model.
Table 2:
Gateway Usable Customer
Customer IP Subnet Address Hosts Hosts
======== ============ ======= ====== ========
A 1.1.1.0/24 1.1.1.1 10 .2-.11
B 1.1.1.0/24 1.1.1.1 5 .12-.16
C 1.1.1.0/24 1.1.1.1 1 .17
Customer A's initial deployment consists of 2 hosts, though they
project growth of up to 10 hosts. As a result, they're allocated
McPherson, Dykes [Page 4]
INTERNET DRAFT December 2000
the IP address range 1.1.1.2 - 1.1.1.11. The gateway address for the
customer is 1.1.1.1, the subnet is 1.1.1.0/24.
Customer B's initial deployment consists of 2 hosts, though they
project growth of up to 5 hosts. As a result, they're allocated the
IP address range 1.1.1.12 - 1.1.1.16. The gateway address for the
customer is 1.1.1.1, the subnet is 1.1.1.0/24.
Customer C's initial deployment consists of 1 host, and they have no
plans of deploying additional hosts. As a result, they're allocated
the IP address 1.1.1.17. The gateway address for the customer is
1.1.1.1, the subnet is 1.1.1.0/24.
The sum of address requirements for all three customers is 16. As a
result, only 16 addresses are allocated within the subnet. These 16
addresses, combined with the global default gateway address of
1.1.1.1, as well as the subnetwork number of 1.1.1.0 and directed
broadcast of 1.1.1.255, result in a total of 19 addresses used. This
leaves 236 additional usable hosts address with the IP subnet.
Now, if customer A only grows to use 3 of his available addresses,
the additional IP addresses can be used for other customers.
Also, assume customer C determines the need to deploy one additional
host, and as such, requires one additional IP address. The customer
is simply allocated the next available IP address within the subnet,
their default gateway remains the same.
The benefits of such a model are obvious, especially when employed in
large LANs or MANs.
6. Use of Directed Broadcasts
This specification provides no support for directed broadcasts.
Specifically, the <net, subnet, -1> directed broadcast address can
only apply to one of the Layer 2 broadcast domains.
Though use of directed broadcast is frowned upon in the Internet
today, there remain a number of applications, primarily in the
enterprise arena, that continue to use them. As such, care should be
taken to understand the implications of using these applications in
conjunction with the addressing model outlined in this specification.
McPherson, Dykes [Page 5]
INTERNET DRAFT December 2000
7. Multicast Considerations
It is assumed that the Layer 2 multicast domain will be the same as
the Layer 2 broadcast domain (i.e. VLAN). As such, this means that
for an IP multicast packet to reach all potential receivers in the IP
subnet the multicast router(s) attached to the IP subnet need to
employ something akin to IP host routes for the sender in order for
the Reverse Path Forwarding check to work.
8. Deployment Considerations
Extreme Networks has a working implementation of this model that has
been deployed in service provider data center environments for over a
year now. Other vendors are rumored to be developing similar
functionality.
9. Security Considerations
One obvious issue that does arise with this model is the
vulnerabilities created by permitting arbitrary allocation of
addresses across disparate broadcast domains. It is advised that
address space ranges be made sticky. That is, when an address or
range of addresses is allocated to a given sub-VLAN, reception of IP
or ARP packets on a sub-VLAN with a source IP address that isn't
allocated to the sub-VLAN should be discarded, and perhaps trigger a
logging message or other administrative event.
Implementation details are intentionally omitted as all functions in
this document should remain local to the super-VLAN router. As such,
no interoperability issues with existing protocols should result.
10. Acknowledgements
Thanks to Mike Hollyman and Erik Nordmark for their feedback.
McPherson, Dykes [Page 6]
INTERNET DRAFT December 2000
11. References
[802.1Q] IEEE 802.1Q, "Virtual LANs".
12. Authors' Address
Danny McPherson
Amber Networks, Inc.
2465 Augustine Drive
Santa Clara, CA 95054
Phone: +1 408.486.6336
Email: danny@ambernetworks.com
Barry Dykes
OneSecure, Inc.
2000 S. Colorado Blvd Suite 2-1100
Denver, CO. 80222
Phone: +1 303-691-0104
Email: bdykes@onesecure.com
McPherson, Dykes [Page 7]
[/snip]
------- End of Forwarded Message