|
Previous
|
Content
|
Next
|
|
|
1.3.- Traffic Engineering
in Internet |
|
|
|
|
In term of traffic management, Internet has been a best effort
service environment. Very limited traffic management capabilities exist in
IP networks to provide differentiated queue management and scheduling
service to packets belonging to different classes. |
|
In terms of routing control, Internet has employed distributed
protocols for intra-domain routing. These protocols are highly scalable and
resilient. However, they are based on simple algorithms for path selection
which have very limited functionality to allow flexible control of the path
selection process. |
| Dynamic Routing in the Internet |
|
Routing algorithms used in the Internet are adaptations of the
Shortest Path Tree (SPF) algorithm, where cost are based on link
metrics. These metrics can be static or dynamic quantities. Static metrics
are assigned administratively according to local criteria. Dynamic metrics
are assigned dynamically based on some measure such as delay,
packet loss or throughput. |
| Static link metric assignment can
easily lead to unfavorable scenarios where some links become congested while
some others remain lightly loaded. Even using dynamic link metric assignment
in accordance with the traffic being routed (using a traffic matrix
like Q-OSPF does), unbalanced load in the network
can still occur due to some factors like resources not being deployed in the
most optimal locations, forecasting errors, and dynamics in the traffic
matrix due to the temporal nature of traffic patterns. All these
considerations are moving the interest in path-oriented technologies
such as "explicit routing" and "constraint-based" routing
(implemented by MPLS). |
| ToS Routing |
| ToS Routing involves
different routes going to the same destination with selection dependent upon
the ToS field of an IP packet. The ToS classes may be
classified as low delay and high throughput. Each link is
associated with multiple link costs and each link cost is used to compute
routes for a particular ToS. A separate SPF is computed for
each ToS. The SPF algorithm must be run for each ToS
resulting in very expensive computation. Classical ToS-based routing
is now outdated as the IP header's ToS field has been replaced by an
IP header's DiffServ (DS) field. |
| Effective traffic engineering is
difficult to perform in classical ToS-based routing because each
class still relies exclusively on SPF routing which results in
localization of traffic concentration within the network. |
| Equal Cost Multi-Path (ECMP) |
| ECMP attempts to address the
deficiency in the SPF interior gateway routing systems. ECMP
works so that if two or more equal shortest paths exist between two nodes,
the traffic between the nodes is distributed among the multiple equal-cost
paths. Traffic distribution across the equal-cost paths is doee as follows: |
- Packet-based in a round-robin fashion. This approach can
easily cause out-of-order packets.
- Flow-based using hashing on source and destination. This
approach will depend upon the number and distribution of flows, being more
unpredictable in enterprise networks where the number of flows is
relatively small and less heterogeneous, but it is generally effective in
core public networks where the number of flows is large and heterogeneous.
|
|
|
|
| In ECMP, link costs are
static and bandwidth constraint are not considered, so it tries to
distribute the traffic as equally as possible among the different paths
ignoring the congestion status of each of them. As a result, some paths
could be more congested than others. Another drawback is that load sharing
cannot be achieved on paths having non-identical costs. |
| Overlay Model |
| In the overlay model, a
virtual-circuit network, such as ATM, FR, or WDM,
provides connectivity between routers that are located at the edges of a
VC cloud. In this mode, two routers that are connected through a VC
see a direct adjacency between themselve, being this independent of
the VC's physical route. This way, the model decouples the logical
topology from the physical topology that the VC network manages. |
| The overlay model based on ATM
or FR, enables the employing of traffic engineering concepts to
perform path optimization by rearranging the VCs, so that a VC
on a congested or sub-optimal physical link can be re-routed to a less
congested or more optimal one. Traffic engineering (TE) is also employed to
establish relationships between the traffic management parameters (e.g.,
PCR, SCR, and MBS for ATM) of the VC
technology and the current traffic traversing each VC. |
| The overlay model requires the
management of two separate networks with different technologies (e.g., IP
and ATM), resulting in increased operational complexity and cost. In
the fully-meshed overlay model, each router would peer to every other
router in the network, resulting in a quadratic function between the number
of adjacencies and the number of routers. |
| Constraint-Based Routing (CBR) |
| CBR refers to a class of
routing system that compute routes through a network subject to the
satisfaction of a set of constraints and requeriments imposed by the network
itself or by administrative policies. CBR tries to optimize network
performance while minimizing costs. |
| Constraints may include
bandwidth, hop count, delay and policy instruments
such as resource class attributes, or domain specific attributes
based on the network technology, which impose restrictions on the solution
space of the routing function. Path-oriented tecnologies such as
MPLS have made CBR feasible and attractive in public IP
networks. |
| Unlike QoS routing, which
addresses the issue of routing individual traffic flows to satisfy
prescribed flow based QoS requeriments subject to network resource
availability, CBR is applicable to traffic aggregate as well as
flows, and may be subject to a wide variety of constraints which may include
policy restrictions. |
|
|
Previous
|
Content
|
Next
|