1 /*! \page platform Describing the virtual platform
5 In order to run any simulation, SimGrid must be provided with three things:
6 something to run (i.e., your code), a description of the platform on which you want to simulate your application, and
7 information about the deployment of the application: Which process should be executed onto which processor/core?
9 For the last two items, there are essentially three possible ways you can provide
10 this information as an input:
11 \li You can program, if you're using MSG, some of the platform and
12 deployment functions. If you choose to follow this approach, check the dedicated documentation
13 (\ref msg_simulation).
14 \li You can use two XML files: one for the platform description and the other for the deployment.
15 \li You can program the description of your platform in Lua format.
17 For more information on SimGrid's deployment features, please refer to the \ref deployment section.
19 The platform description may be intricate. This documentation is all about how to write this file. First, the basic
20 concepts are introduced. Then, advanced options are explained. Finally, some hints and tips on how to write a better
21 platform description are given.
23 \section pf_overview Some words about XML and DTD
25 We opted for XML not only because it is extensible but also because many tools (and plugins for existing tools) are
26 available that facilitate editing and validating XML files. Furthermore, libraries that parse XML are often already
27 available and very well tested.
29 The XML checking is done based on the [simgrid.dtd](http://simgrid.gforge.inria.fr/simgrid/simgrid.dtd) Document Type
30 Definition (DTD) file.
32 If you read the DTD, you should notice the following:
33 \li The platform tag has a version attribute. The current version is <b>4</b>. This attribute might be used in the
34 provide backward compatibility.
35 \li The DTD contains definitions for both the platform description and deployment files used by SimGrid.
37 \section pf_basics Basic concepts
39 Nowadays, the Internet is composed of a bunch of independently managed
40 networks. Within each of those networks, there are entry and exit
41 points (most of the time, you can both enter and exit through the same
42 point); this allows to leave the current network and reach other
43 networks, possibly even in other locations.
44 At the upper level, such a network is called
45 <b>Autonomous System (AS)</b>, while at the lower level it is named
46 sub-network, or LAN (local area network).
47 They are indeed autonomous: routing is defined
48 (within the limits of his network) by the administrator, and so, those
49 networks can operate without a connection to other
50 networks. So-called gateways allow you to go from one network to
51 another, if such a (physical) connection exists. Every node in one network
52 that can be directly reached (i.e., without traversing other nodes) from
53 another network is called a gateway.
54 Each autonomous system consists of equipment such as cables (network links),
55 routers and switches as well as computers.
57 The structure of the SimGrid platform description relies exactly on the same
58 concept as a real-life platform (see above). Every resource (computers,
59 network equipment etc.) belongs to an AS, which can be defined by using the
60 \<AS\> tag. Within an AS, the routing between its elements can be defined
61 abitrarily. There are several modes for routing, and exactly one mode must be
62 selected by specifying the routing attribute in the AS tag:
65 <AS id="AS0" routing="Full">
69 Other supported values for the routing attribute can be found below, Section
72 There is also the ``<route>`` tag; this tag takes two attributes, ``src`` (source)
73 and ``dst`` (destination). Both source and destination must be valid identifiers
74 for routers (these will be introduced later). Contained by the ``<route>`` are
75 network links; these links must be used in order to communicate from the source
76 to the destination specified in the tag. Hence, a route merely describes
77 how to reach a router from another router.
80 More information and (code-)examples can be found in Section \ref pf_rm.
82 An AS can also contain itself one or more AS; this allows you to
83 model the hierarchy of your platform.
85 ### Within each AS, the following types of resources exist:
87 %Resource | Documented in Section | Description
88 --------------- | --------------------- | -----------
89 AS | | Every Autonomous System (AS) may contain one or more AS.
90 host | \ref pf_host | This entity carries out the actual computation. For this reason, it contains processors (with potentially multiple cores).
91 router | \ref pf_router | In SimGrid, routers are used to provide helpful information to routing algorithms. Routers may also act as gateways, connecting several autonomous systems with each other.
92 link | \ref pf_link | In SimGrid, (network)links define a connection between two or potentially even more resources. Every link has a bandwidth and a latency and may potentially experience congestion.
93 cluster | \ref pf_cluster | In SimGrid, clusters were introduced to model large and homogenous environments. They are not really a resource by themselves - technically, they are only a shortcut, as they will internally set up all the hosts, network and routing for you, i.e., using this resource, one can easily setup thousands of hosts and links in a few lines of code. Each cluster is itself an AS.
95 %As it is desirable to interconnect these resources, a routing has to be
96 defined. The AS is supposed to be Autonomous, hence this has to be done at the
97 AS level. The AS handles two different types of entities (<b>host/router</b>
98 and <b>AS</b>). However, the user is responsible to define routes between those resources,
99 otherwise entities will be unconnected and therefore unreachable from other
100 entities. Although several routing algorithms are built into SimGrid (see
101 \ref pf_rm), you might encounter a case where you want to define routes
102 manually (for instance, due to specific requirements of your platform).
104 There are three tags to use:
105 \li <b>ASroute</b>: to define routes between two <b>AS</b>
106 \li <b>route</b>: to define routes between two <b>host/router</b>
107 \li <b>bypassRoute</b>: to define routes between two <b>AS</b> that
108 will bypass default routing (as specified by the ``routing`` attribute
109 supplied to ``<AS>``, see above).
111 Here is an illustration of these concepts:
113 
114 Circles represent processing units and squares represent network routers. Bold
115 lines represent communication links. AS2 models the core of a national
116 network interconnecting a small flat cluster (AS4) and a larger
117 hierarchical cluster (AS5), a subset of a LAN (AS6), and a set of peers
118 scattered around the world (AS7).
120 \section pf_pftags Resource description
122 \subsection pf_As Platform: The <AS> tag
124 The concept of an AS was already outlined above (Section \ref pf_basics);
125 recall that the AS is so important because it groups other resources (such
126 as routers/hosts) together (in fact, these resources must be contained by
129 Available attributes :
131 Attribute name | Mandatory | Values | Description
132 --------------- | --------- | ------ | -----------
133 id | yes | String | The identifier of an AS; facilitates referring to this AS. ID must be unique.
134 routing | yes | Full\| Floyd\| Dijkstra\| DijkstraCache\| None\| Vivaldi\| Cluster | See Section \ref pf_rm for details.
139 <AS id="AS0" routing="Full">
140 <host id="host1" power="1000000000"/>
141 <host id="host2" power="1000000000"/>
142 <link id="link1" bandwidth="125000000" latency="0.000100"/>
143 <route src="host1" dst="host2"><link_ctn id="link1"/></route>
147 In this example, AS0 contains two hosts (host1 and host2). The route
148 between the hosts goes through link1.
150 \subsection pf_Cr Computing resources: hosts, clusters and peers.
152 \subsubsection pf_host <host/>
154 A <b>host</b> represents a computer/node card. Every host is able to execute
155 code and it can send and receive data to/from other hosts. Most importantly,
156 a host can contain more than 1 core.
160 Attribute name | Mandatory | Values | Description
161 --------------- | --------- | ------ | -----------
162 id | yes | String | The identifier of the host. facilitates referring to this AS.
163 power | yes | double (must be > 0.0) | Computational power of every core of this host in FLOPS. Must be larger than 0.0.
164 core | no | int (Default: 1) | The number of cores of this host. If more than one core is specified, the "power" parameter refers to every core, i.e., the total computational power is no_of_cores*power.<br /> If 6 cores are specified, up to 6 tasks can be executed without sharing the computational power; if more than 6 tasks are executed, computational power will be shared among these tasks. <br /> <b>Warning:</b> Although functional, this model was never scientifically assessed.
165 availability | no | int | <b>Specify if the percentage of power available.</b> (What? TODO)
166 availability_file| no | string | (Relative or absolute) filename to use as input; must contain availability traces for this host. The syntax of this file is defined below. <br /> <b>Note:</b> The filename must be specified with your system's format.
167 state | no | ON\|OFF<br/> (Default: ON) | Is this host running or not?
168 state_file | no | string | Same mechanism as availability_file.<br /> <b>Note:</b> The filename must be specified with your system's format.
169 coordinates | no | string | Must be provided when choosing the Vivaldi, coordinate-based routing model for the AS the host belongs to. More details can be found in the Section \ref pf_P2P_tags.
171 ### Possible children: ###
173 Tag name | Description | Documentation
174 ------------ | ----------- | -------------
175 \<mount/\> | Defines mounting points between some storage resource and the host. | \ref pf_storage_entity_mount
176 \<prop/\> | The prop tag allows you to define additional information on this host following the attribute/value schema. You may want to use it to give information to the tool you use for rendering your simulation, for example. | N/A
181 <host id="host1" power="1000000000"/>
182 <host id="host2" power="1000000000">
183 <prop id="color" value="blue"/>
184 <prop id="rendershape" value="square"/>
189 \anchor pf_host_dynamism
190 ### Expressing dynamism ###
192 SimGrid provides mechanisms to change a hosts' availability over
193 time, using the ``availability_file`` attribute to the ``\<host\>`` tag
194 and a separate text file whose syntax is exemplified below.
196 #### Adding a trace file ####
199 <platform version="4">
200 <host id="bob" power="500Gf" availability_file="bob.trace" />
204 #### Example of "bob.trace" file ####
213 Let us begin to explain this example by looking at line 2. (Line 1 will become clear soon).
214 The first column describes points in time, in this case, time 0. The second column
215 describes the relative amount of power this host is able to deliver (relative
216 to the maximum performance specified in the ``\<host\>`` tag). (Clearly, the
217 second column needs to contain values that are not smaller than 0 and not larger than 1).
218 In this example, our host will deliver 500 Mflop/s at time 0, as 500 Mflop/s is the
219 maximum performance of this host. At time 11.0, it will
220 deliver half of its maximum performance, i.e., 250 Mflop/s until time 20.0 when it will
221 will start delivering 80\% of its power. In this example, this amounts to 400 Mflop/s.
223 Since the periodicity in line 1 was set to be 1.0, i.e., 1 timestep, this host will
224 continue to provide 500 Mflop/s from time 21. From time 32 it will provide 250 MFlop/s and so on.
226 ### Changing initial state ###
228 It is also possible to specify whether the host is up or down by setting the
229 ``state`` attribute to either <b>ON</b> (default value) or <b>OFF</b>.
231 #### Example: Expliciting the default value "ON" ####
234 <platform version="4">
235 <host id="bob" power="500Gf" state="ON" />
239 If you want this host to be unavailable, simply substitute ON with OFF.
241 \anchor pf_host_churn
242 ### Expressing churn ###
244 To express the fact that a host can change state over time (as in P2P
245 systems, for instance), it is possible to use a file describing the time
246 at which the host is turned on or off. An example of the content
247 of such a file is presented below.
249 #### Adding a state file ####
252 <platform version="4">
253 <host id="bob" power="500Gf" state_file="bob.fail" />
257 #### Example of "bob.fail" file ####
265 A zero value means <b>down</b> (i.e., OFF) while a positive one means <b>up and
266 running</b> (i.e., ON). From time 0.0 to time 1.0, the host is on as usual. At time 1.0, it is
267 turned off and at time 2.0, it is turned on again until time 12 (2 plus the
268 periodicity 10). It will be turned off again at time 13.0 until time 23.0, and
272 \subsubsection pf_cluster <cluster>
274 ``<cluster />`` represents a machine-cluster. It is most commonly used
275 when one wants to define many hosts and a network quickly. Technically,
276 ``cluster`` is a meta-tag: <b>from the inner SimGrid point of
277 view, a cluster is an AS where some optimized routing is defined</b>.
278 The default inner organization of the cluster is as follow:
284 ____________|__________|_____________ backbone
286 l0| l1| l2| l97| l96 | | l99
292 Here, a set of <b>host</b>s is defined. Each of them has a <b>link</b>
293 to a central backbone (backbone is a link itself, as a link can
294 be used to represent a switch, see the switch / link section
295 below for more details about it). A <b>router</b> allows to connect a
296 <b>cluster</b> to the outside world. Internally,
297 SimGrid treats a cluster as an AS containing all hosts: the router is the default
298 gateway for the cluster.
300 There is an alternative organization, which is as follows:
314 The principle is the same, except that there is no backbone. This representation
315 can be obtained easily: just do not set the bb_* attributes.
318 Attribute name | Mandatory | Values | Description
319 --------------- | --------- | ------ | -----------
320 id | yes | string | The identifier of the cluster. Facilitates referring to this cluster.
321 prefix | yes | string | Each node of the cluster has to have a name. This name will be prefixed with this prefix.
322 suffix | yes | string | Each node of the cluster will be suffixed with this suffix
323 radical | yes | string | Regexp used to generate cluster nodes name. Syntax: "10-20" will give you 11 machines numbered from 10 to 20, "10-20;2" will give you 12 machines, one with the number 2, others numbered as before. The produced number is concatenated between prefix and suffix to form machine names.
324 power | yes | int | Same as the ``power`` attribute of the ``\<host\>`` tag.
325 core | no | int (default: 1) | Same as the ``core`` attribute of the ``\<host\>`` tag.
326 bw | yes | int | Bandwidth for the links between nodes and backbone (if any). See the \ref pf_link "link section" for syntax/details.
327 lat | yes | int | Latency for the links between nodes and backbone (if any). See <b>link</b> section for syntax/details.
328 sharing_policy | no | string | Sharing policy for the links between nodes and backbone (if any). See <b>link</b> section for syntax/details.
329 bb_bw | no | int | Bandwidth for backbone (if any). See <b>link</b> section for syntax/details. If bb_bw and bb_lat (see below) attributes are omitted, no backbone is created (alternative cluster architecture <b>described before</b>).
330 bb_lat | no | int | Latency for backbone (if any). See <b>link</b> section for syntax/details. If bb_lat and bb_bw (see above) attributes are omitted, no backbone is created (alternative cluster architecture <b>described before</b>).
331 bb_sharing_policy | no | string | Sharing policy for the backbone (if any). See <b>link</b> section for syntax/details.
332 limiter_link | no | int | Bandwidth for limiter link (if any). This adds a specific link for each node, to set the maximum bandwidth reached when communicating in both directions at the same time. In theory this value should be 2*bw for fullduplex links, but in reality this might be less. This value will depend heavily on the communication model, and on the cluster's hardware, so no default value can be set, this has to be measured. More details can be obtained in <a href="https://hal.inria.fr/hal-00919507/"> "Toward Better Simulation of MPI Applications on Ethernet/TCP Networks"</a>
333 loopback_bw | no | int | Bandwidth for loopback (if any). See <b>link</b> section for syntax/details. If loopback_bw and loopback_lat (see below) attributes are omitted, no loopback link is created and all intra-node communication will use the main network link of the node. Loopback link is a \ref pf_sharing_policy_fatpipe "\b FATPIPE".
334 loopback_lat | no | int | Latency for loopback (if any). See <b>link</b> section for syntax/details. See loopback_bw for more info.
335 topology | no | FLAT\|TORUS\|FAT_TREE (default: FLAT) | Network topology to use. SimGrid currently supports FLAT (with or without backbone, as described before), <a href="http://en.wikipedia.org/wiki/Torus_interconnect">TORUS </a> and FAT_TREE attributes for this tag.
336 topo_parameters | no | string | Specific parameters to pass for the topology defined in the topology tag. For torus networks, comma-separated list of the number of nodes in each dimension of the torus. For fat trees, refer to \ref simgrid::surf::AsClusterFatTree "AsClusterFatTree documentation".
340 the router name is defined as the resulting String in the following
344 router_name = prefix + clusterId + _router + suffix;
348 #### Cluster example ####
350 Consider the following two (and independent) uses of the ``cluster`` tag:
353 <cluster id="my_cluster_1" prefix="" suffix="" radical="0-262144"
354 power="1e9" bw="125e6" lat="5E-5"/>
356 <cluster id="my_cluster_2" prefix="c-" suffix=".me" radical="0-99"
357 power="1e9" bw="125e6" lat="5E-5"
358 bb_bw="2.25e9" bb_lat="5E-4"/>
361 The second example creates one router and 100 machines with the following names:
363 c-my_cluster_2_router.me
371 \subsubsection pf_cabinet <cabinet>
374 This tag is only available when the routing mode of the AS
375 is set to ``Cluster``.
377 The ``<cabinet />`` tag is, like the \ref pf_cluster "<cluster>" tag,
378 a meta-tag. This means that it is simply a shortcut for creating a set of (homogenous) hosts and links quickly;
379 unsurprisingly, this tag was introduced to setup cabinets in data centers quickly. Unlike
380 <cluster>, however, the <cabinet> assumes that you create the backbone
381 and routers yourself; see our examples below.
385 Attribute name | Mandatory | Values | Description
386 --------------- | --------- | ------ | -----------
387 id | yes | string | The identifier of the cabinet. Facilitates referring to this cluster.
388 prefix | yes | string | Each node of the cabinet has to have a name. This name will be prefixed with this prefix.
389 suffix | yes | string | Each node of the cabinet will be suffixed with this suffix
390 radical | yes | string | Regexp used to generate cabinet nodes name. Syntax: "10-20" will give you 11 machines numbered from 10 to 20, "10-20;2" will give you 12 machines, one with the number 2, others numbered as before. The produced number is concatenated between prefix and suffix to form machine names.
391 power | yes | int | Same as the ``power`` attribute of the \ref pf_host "<host>" tag.
392 bw | yes | int | Bandwidth for the links between nodes and backbone (if any). See the \ref pf_link "link section" for syntax/details.
393 lat | yes | int | Latency for the links between nodes and backbone (if any). See the \ref pf_link "link section" for syntax/details.
396 Please note that as of now, it is impossible to change attributes such as,
397 amount of cores (always set to 1), the initial state of hosts/links
398 (always set to ON), the sharing policy of the links (always set to \ref pf_sharing_policy_fullduplex "FULLDUPLEX").
402 The following example was taken from ``examples/platforms/meta_cluster.xml`` and
403 shows how to use the cabinet tag.
406 <AS id="my_cluster1" routing="Cluster">
407 <cabinet id="cabinet1" prefix="host-" suffix=".cluster1"
408 power="1Gf" bw="125MBps" lat="100us" radical="1-10"/>
409 <cabinet id="cabinet2" prefix="host-" suffix=".cluster1"
410 power="1Gf" bw="125MBps" lat="100us" radical="11-20"/>
411 <cabinet id="cabinet3" prefix="host-" suffix=".cluster1"
412 power="1Gf" bw="125MBps" lat="100us" radical="21-30"/>
414 <backbone id="backbone1" bandwidth="2.25GBps" latency="500us"/>
419 Please note that you must specify the \ref pf_backbone "<backbone>"
420 tag by yourself; this is not done automatically and there are no checks
421 that ensure this backbone was defined.
423 The hosts generated in the above example are named host-1.cluster, host-2.cluster1
427 \subsubsection pf_peer The <peer> tag
429 This tag represents a peer, as in Peer-to-Peer (P2P) networks. However, internally,
430 SimGrid transforms a peer into an AS (similar to Cluster). Hence, this tag
431 is virtually only a shortcut that comes with some pre-defined resources
432 and values. These are:
434 \li A tiny AS whose routing type is cluster is created
436 \li Two links: One for download and one for upload. This is
437 convenient to use and simulate stuff under the last mile model (e.g., ADSL peers).
438 \li It connects the two links to the host
439 \li It creates a router (a gateway) that serves as an entry point for this peer zone.
440 This router has coordinates.
444 Attribute name | Mandatory | Values | Description
445 --------------- | --------- | ------ | -----------
446 id | yes | string | The identifier of the peer. Facilitates referring to this peer.
447 power | yes | int | See the description of the ``host`` tag for this attribute
448 bw_in | yes | int | Bandwidth downstream
449 bw_out | yes | int | Bandwidth upstream
450 lat | yes | double | Latency for both up- and downstream, in seconds.
451 coordinates | no | string | Coordinates of the gateway for this peer. Example value: 12.8 14.4 6.4
452 sharing_policy | no | SHARED\|FULLDUPLEX (default: FULLDUPLEX) | Sharing policy for links. See <b>link</b> description for details.
453 availability_file| no | string | Availability file for the peer. Same as host availability file. See <b>host</b> description for details.
454 state_file | no | string | State file for the peer. Same as host state file. See <b>host</b> description for details.
456 Internally, SimGrid transforms any ``\<peer/\>`` construct such as
459 coordinates="12.8 14.4 6.4"
465 into an ``\<AS\>`` (see Sections \ref pf_basics and \ref pf_As). In fact, this example of the ``\<peer/\>`` tag
466 is completely equivalent to the following declaration:
469 <AS id="as_FOO" routing="Cluster">
470 <host id="peer_FOO" power="1.5Gf"/>
471 <link id="link_FOO_UP" bandwidth="2.25GBps" latency="500us"/>
472 <link id="link_FOO_DOWN" bandwidth="2.25GBps" latency="500us"/>
473 <router id="router_FOO" coordinates="25.5 9.4 1.4"/>
474 <host_link id="peer_FOO" up="link_FOO_UP" down="link_FOO_DOWN"/>
479 \subsection pf_ne Network equipments: links and routers
481 There are two tags at all times available to represent network entities and
482 several other tags that are available only in certain contexts.
483 1. ``<link>``: Represents a entity that has a limited bandwidth, a
484 latency, and that can be shared according to TCP way to share this
487 The concept of links in SimGrid may not be intuitive, as links are not
488 limited to connecting (exactly) two entities; in fact, you can have more than
489 two equipments connected to it. (In graph theoretical terms: A link in
490 SimGrid is not an edge, but a hyperedge)
492 2. ``<router/>``: Represents an entity that a message can be routed
493 to, but that is unable to execute any code. In SimGrid, routers have also
494 no impact on the performance: Routers do not limit any bandwidth nor
495 do they increase latency. As a matter of fact, routers are (almost) ignored
496 by the simulator when the simulation has begun.
498 3. ``<backbone/>``: This tag is only available when the containing AS is
499 used as a cluster (i.e., mode="Cluster")
502 If you want to represent an entity like a switch, you must use ``<link>`` (see section). Routers are used
503 to run some routing algorithm and determine routes (see Section \ref pf_routing for details).
505 \subsubsection pf_router <router/>
507 %As said before, <b>router</b> is used only to give some information
508 for routing algorithms. So, it does not have any attributes except :
512 Attribute name | Mandatory | Values | Description
513 --------------- | --------- | ------ | -----------
514 id | yes | string | The identifier of the router to be used when referring to it.
515 coordinates | no | string | Must be provided when choosing the Vivaldi, coordinate-based routing model for the AS the router belongs to. More details can be found in the Section \ref pf_P2P_tags.
520 <router id="gw_dc1_horizdist"/>
523 \subsubsection pf_link <link/>
525 Network links can represent one-hop network connections. They are
526 characterized by their id and their bandwidth; links can (but may not) be subject
531 Attribute name | Mandatory | Values | Description
532 --------------- | --------- | ------ | -----------
533 id | yes | string | The identifier of the link to be used when referring to it.
534 bandwidth | yes | int | Maximum bandwidth for this link, given in bytes/s
535 latency | no | double (default: 0.0) | Latency for this link.
536 sharing_policy | no | \ref sharing_policy_shared "SHARED"\|\ref pf_sharing_policy_fatpipe "FATPIPE"\|\ref pf_sharing_policy_fullduplex "FULLDUPLEX" (default: SHARED) | Sharing policy for the link.
537 state | no | ON\|OFF (default: ON) | Allows you to to turn this link on or off (working / not working)
538 bandwidth_file | no | string | Allows you to use a file as input for bandwidth.
539 latency_file | no | string | Allows you to use a file as input for latency.
540 state_file | no | string | Allows you to use a file as input for states.
543 #### Possible shortcuts for ``latency`` ####
545 When using the latency attribute, you can specify the latency by using the scientific
546 notation or by using common abbreviations. For instance, the following three tags
550 <link id="LINK1" bandwidth="125000000" latency="5E-6"/>
551 <link id="LINK1" bandwidth="125000000" latency="5us"/>
552 <link id="LINK1" bandwidth="125000000" latency="0.000005"/>
555 Here, the second tag uses "us", meaning "microseconds". Other shortcuts are:
557 Name | Abbreviation | Time (in seconds)
558 ---- | ------------ | -----------------
559 Week | w | 7 * 24 * 60 * 60
560 Day | d | 24 * 60 * 60
564 Millisecond | ms | 0.001 = 10^(-3)
565 Microsecond | us | 0.000001 = 10^(-6)
566 Nanosecond | ns | 0.000000001 = 10^(-9)
567 Picosecond | ps | 0.000000000001 = 10^(-12)
569 #### Sharing policy ####
571 \anchor sharing_policy_shared
572 By default a network link is \b SHARED, i.e., if two or more data flows go
573 through a link, the bandwidth is shared fairly among all data flows. This
574 is similar to the sharing policy TCP uses.
576 \anchor pf_sharing_policy_fatpipe
577 On the other hand, if a link is defined as a \b FATPIPE,
578 each flow going through this link will be provided with the complete bandwidth,
579 i.e., no sharing occurs and the bandwidth is only limiting each flow individually.
580 Please note that this is really on a per-flow basis, not only on a per-host basis!
581 The complete bandwidth provided by this link in this mode
582 is ``number_of_flows*bandwidth``, with at most ``bandwidth`` being available per flow.
584 Using the FATPIPE mode allows to model backbones that won't affect performance
587 \anchor pf_sharing_policy_fullduplex
588 The last mode available is \b FULLDUPLEX. This means that SimGrid will
589 automatically generate two links (one carrying the suffix _UP and the other the
590 suffix _DOWN) for each ``<link>`` tag. This models situations when the direction
591 of traffic is important.
594 Transfers from one side to the other will interact similarly as
595 TCP when ACK returning packets circulate on the other direction. More
596 discussion about it is available in the description of link_ctn description.
598 In other words: The SHARED policy defines a physical limit for the bandwidth.
599 The FATPIPE mode defines a limit for each application,
600 with no upper total limit.
603 Tip: By using the FATPIPE mode, you can model big backbones that
604 won't affect performance (except latency).
609 <link id="SWITCH" bandwidth="125000000" latency="5E-5" sharing_policy="FATPIPE" />
612 #### Expressing dynamism and failures ####
614 Similar to hosts, it is possible to declare links whose state, bandwidth
615 or latency changes over time (see Section \ref pf_host_dynamism for details).
617 In the case of network links, the ``bandwidth`` and ``latency`` attributes are
618 replaced by the ``bandwidth_file`` and ``latency_file`` attributes.
619 The following XML snippet demonstrates how to use this feature in the platform
620 file. The structure of the files "link1.bw" and "link1.lat" is shown below.
623 <link id="LINK1" state_file="link1.fail" bandwidth="80000000" latency=".0001" bandwidth_file="link1.bw" latency_file="link1.lat" />
627 Even if the syntax is the same, the semantic of bandwidth and latency
628 trace files differs from that of host availability files. For bandwidth and
629 latency, the corresponding files do not
630 express availability as a fraction of the available capacity but directly in
631 bytes per seconds for the bandwidth and in seconds for the latency. This is
632 because most tools allowing to capture traces on real platforms (such as NWS)
633 express their results this way.
635 ##### Example of "link1.bw" file #####
643 In this example, the bandwidth changes repeatedly, with all changes
644 being repeated every 12 seconds.
646 At the beginning of the the simulation, the link's bandwidth is 80,000,000
647 B/s (i.e., 80 Mb/s); this value was defined in the XML snippet above.
648 After four seconds, it drops to 40 Mb/s (line 2), and climbs
649 back to 60 Mb/s after another 4 seconds (line 3). The value does not change any
650 more until the end of the period, that is, after 12 seconds have been simulated).
651 At this point, periodicity kicks in and this behavior is repeated: Seconds
652 12-16 will experience 80 Mb/s, 16-20 40 Mb/s etc.).
654 ##### Example of "link1.lat" file #####
663 In this example, the latency varies with a period of 5 seconds.
664 In the xml snippet above, the latency is initialized to be 0.0001s (100µs). This
665 value will be kept during the first second, since the latency_file contains
666 changes to this value at second one, two and three.
667 At second one, the value will be 0.001, i.e., 1ms. One second later it will
668 be adjusted to 0.01 (or 10ms) and one second later it will be set again to 1ms. The
669 value will not change until second 5, when the periodicity defined in line 1
670 kicks in. It then loops back, starting at 100µs (the initial value) for one second.
673 #### The ``<prop/>`` tag ####
675 Similar to the ``<host>`` tag, a link may also contain the ``<prop/>`` tag; see the host
676 documentation (Section \ref pf_host) for an example.
679 \subsubsection pf_backbone <backbone/>
682 This tag is <b>only available</b> when the containing AS uses the "Cluster" routing mode!
684 Using this tag, you can designate an already existing link to be a backbone.
686 Attribute name | Mandatory | Values | Description
687 --------------- | --------- | ------ | -----------
688 id | yes | string | Name of the link that is supposed to act as a backbone.
690 \subsection pf_storage Storage
693 This is a prototype version that should evolve quickly, hence this
694 is just some doc valuable only at the time of writing.
695 This section describes the storage management under SimGrid ; nowadays
696 it's only usable with MSG. It relies basically on linux-like concepts.
697 You also may want to have a look to its corresponding section in
698 @ref msg_file ; access functions are organized as a POSIX-like
701 \subsubsection pf_sto_conc Storage - Main Concepts
703 The storage facilities implemented in SimGrid help to model (and account for)
704 storage devices, such as tapes, hard-drives, CD or DVD devices etc.
705 A typical situation is depicted in the figure below:
707 \image html ./webcruft/storage_sample_scenario.png
708 \image latex ./webcruft/storage_sample_scenario.png "storage_sample_scenario" width=\textwidth
710 In this figure, two hosts called Bob and Alice are interconnected via a network
711 and each host is physically attached to a disk; it is not only possible for each host to
712 mount the disk they are attached to directly, but they can also mount disks
713 that are in a remote location. In this example, Bob mounts Alice's disk remotely
714 and accesses the storage via the network.
716 SimGrid provides 3 different entities that can be used to model setups
717 that include storage facilities:
719 Entity name | Description
720 --------------- | -----------
721 \ref pf_storage_entity_storage_type "storage_type" | Defines a template for a particular kind of storage (such as a hard-drive) and specifies important features of the storage, such as capacity, performance (read/write), contents, ... Different models of hard-drives use different storage_types (because the difference between an SSD and an HDD does matter), as they differ in some specifications (e.g., different sizes or read/write performance).
722 \ref pf_storage_entity_storage "storage" | Defines an actual instance of a storage type (disk, RAM, ...); uses a ``storage_type`` template (see line above) so that you don't need to re-specify the same details over and over again.
723 \ref pf_storage_entity_mount "mount" | Must be wrapped by a \ref pf_host tag; declares which storage(s) this host has mounted and where (i.e., the mountpoint).
726 \anchor pf_storage_content_file
727 ### %Storage Content File ###
729 In order to assess exactly how much time is spent reading from the storage,
730 SimGrid needs to know what is stored on the storage device (identified by distinct (file-)name, like in a file system)
731 and what size this content has.
734 The content file is never changed by the simulation; it is parsed once
735 per simulation and kept in memory afterwards. When the content of the
736 storage changes, only the internal SimGrid data structures change.
738 \anchor pf_storage_content_file_structure
739 #### Structure of a %Storage Content File ####
741 Here is an excerpt from two storage content file; if you want to see the whole file, check
742 the file ``examples/platforms/content/storage_content.txt`` that comes with the
745 SimGrid essentially supports two different formats: UNIX-style filepaths should
746 follow the well known format:
749 /lib/libsimgrid.so.3.6.2 12710497
753 /bin/simgrid_update_xml 5018
754 /bin/graphicator 66986
755 /bin/simgrid-colorizer 2993
760 Windows filepaths, unsurprisingly, use the windows style:
763 \Windows\avastSS.scr 41664
764 \Windows\bfsvc.exe 75264
765 \Windows\bootstat.dat 67584
766 \Windows\CoreSingleLanguage.xml 31497
768 \Windows\dchcfg64.exe 335464
769 \Windows\dcmdev64.exe 93288
773 The different file formats come at a cost; in version 3.12 (and most likely
774 in later versions, too), copying files from windows-style storages to unix-style
775 storages (and vice versa) is not supported.
777 \anchor pf_storage_content_file_create
778 #### Generate a %Storage Content File ####
780 If you want to generate a storage content file based on your own filesystem (or at least a filesystem you have access to),
781 try running this command (works only on unix systems):
784 find . -type f -exec ls -1s --block=1 {} \; 2>/dev/null | awk '{ print $2 " " $1}' > ./content.txt
787 \subsubsection pf_storage_entities The Storage Entities
789 These are the entities that you can use in your platform files to include
790 storage in your model. See also the list of our \ref pf_storage_example_files "example files";
791 these might also help you to get started.
793 \anchor pf_storage_entity_storage_type
794 #### \<storage_type\> ####
796 Attribute name | Mandatory | Values | Description
797 --------------- | --------- | ------ | -----------
798 id | yes | string | Identifier of this storage_type; used when referring to it
799 model | yes | string | For reasons of future backwards compatibility only; specifies the name of the model for the storage that should be used
800 size | yes | string | Specifies the amount of available storage space; you can specify storage like "500GiB" or "500GB" if you want. (TODO add a link to all the available abbreviations)
801 content | yes | string | Path to a \ref pf_storage_content_file "Storage Content File" on your system. This file must exist.
802 content_type | no | ("txt_unix"\|"txt_win") | Determines which kind of filesystem you're using; make sure the filenames (stored in that file, see \ref pf_storage_content_file_structure "Storage Content File Structure"!)
804 This tag must contain some predefined model properties, specified via the <model_prop> tag. Here is a list,
805 see below for an example:
807 Property id | Mandatory | Values | Description
808 --------------- | --------- | ------ | -----------
809 Bwrite | yes | string | Bandwidth for write access; in B/s (but you can also specify e.g. "30MBps")
810 Bread | yes | string | Bandwidth for read access; in B/s (but you can also specify e.g. "30MBps")
811 Bconnexion | yes | string | Throughput (of the storage connector) in B/s.
814 A storage_type can also contain the <b><prop></b> tag. The <prop> tag allows you
815 to associate additional information to this <storage_type> and follows the
816 attribute/value schema; see the example below. You may want to use it to give information to
817 the tool you use for rendering your simulation, for example.
819 Here is a complete example for the ``storage_type`` tag:
821 <storage_type id="single_HDD" model="linear_no_lat" size="4000" content_type="txt_unix">
822 <model_prop id="Bwrite" value="30MBps" />
823 <model_prop id="Bread" value="100MBps" />
824 <model_prop id="Bconnection" value="150MBps" />
825 <prop id="Brand" value="Western Digital" />
829 \anchor pf_storage_entity_storage
830 #### <storage> ####
832 ``storage`` attributes:
834 Attribute name | Mandatory | Values | Description
835 -------------- | --------- | ------ | -----------
836 id | yes | string | Identifier of this ``storage``; used when referring to it
837 typeId | yes | string | Here you need to refer to an already existing \ref pf_storage_entity_storage_type "\<storage_type\>"; the storage entity defined by this tag will then inherit the properties defined there.
838 attach | yes | string | Name of a host (see Section \ref pf_host) to which this storage is <i>physically</i> attached to (e.g., a hard drive in a computer)
839 content | no | string | When specified, overwrites the content attribute of \ref pf_storage_entity_storage_type "\<storage_type\>"
840 content_type | no | string | When specified, overwrites the content_type attribute of \ref pf_storage_entity_storage_type "\<storage_type\>"
842 Here are two examples:
845 <storage id="Disk1" typeId="single_HDD" attach="bob" />
847 <storage id="Disk2" typeId="single_SSD"
848 content="content/win_storage_content.txt"
849 content_type="txt_windows" attach="alice" />
852 The first example is straightforward: A disk is defined and called "Disk1"; it is
853 of type "single_HDD" (shown as an example of \ref pf_storage_entity_storage_type "\<storage_type\>" above) and attached
854 to a host called "bob" (the definition of this host is omitted here).
856 The second storage is called "Disk2", is still of the same type as Disk1 but
857 now specifies a new content file (so the contents will be different from Disk1)
858 and the filesystem uses the windows style; finally, it is attached to a second host,
859 called alice (which is again not defined here).
861 \anchor pf_storage_entity_mount
862 #### <mount> ####
865 | Attribute name | Mandatory | Values | Description |
866 | ---------------- | ----------- | -------- | ------------- |
867 | id | yes | string | Refers to a \ref pf_storage_entity_storage "<storage>" entity that will be mounted on that computer |
868 | name | yes | string | Path/location to/of the logical reference (mount point) of this disk
870 This tag must be enclosed by a \ref pf_host tag. It then specifies where the mountpoint of a given storage device (defined by the ``id`` attribute)
871 is; this location is specified by the ``name`` attribute.
873 Here is a simple example, taken from the file ``examples/platform/storage.xml``:
876 <storage_type id="single_SSD" model="linear_no_lat" size="500GiB">
877 <model_prop id="Bwrite" value="60MBps" />
878 <model_prop id="Bread" value="200MBps" />
879 <model_prop id="Bconnection" value="220MBps" />
882 <storage id="Disk2" typeId="single_SSD"
883 content="content/win_storage_content.txt"
884 content_type="txt_windows" attach="alice" />
885 <storage id="Disk4" typeId="single_SSD"
886 content="content/small_content.txt"
887 content_type="txt_unix" attach="denise"/>
889 <host id="alice" power="1Gf">
890 <mount storageId="Disk2" name="c:"/>
893 <host id="denise" power="1Gf">
894 <mount storageId="Disk2" name="c:"/>
895 <mount storageId="Disk4" name="/home"/>
899 This example is quite interesting, as the same device, called "Disk2", is mounted by
900 two hosts at the same time! Note, however, that the host called ``alice`` is actually
901 attached to this storage, as can be seen in the \ref pf_storage_entity_storage "<storage>"
902 tag. This means that ``denise`` must access this storage through the network, but SimGrid automatically takes
903 care of that for you.
905 Furthermore, this example shows that ``denise`` has mounted two storages with different
906 filesystem types (unix and windows). In general, a host can mount as many storage devices as
910 Again, the difference between ``attach`` and ``mount`` is simply that
911 an attached storage is always physically inside (or connected to) that machine;
912 for instance, a USB stick is attached to one and only one machine (where it's plugged-in)
913 but it can only be mounted on others, as mounted storage can also be a remote location.
915 ###### Example files #####
917 \verbinclude example_filelist_xmltag_mount
919 \subsubsection pf_storage_example_files Example files
921 Several examples were already discussed above; if you're interested in full examples,
922 check the the following platforms:
924 1. ``examples/platforms/storage.xml``
925 2. ``examples/platforms/remote_io.xml``
927 If you're looking for some examplary C code, you may find the source code
928 available in the directory ``examples/msg/io/`` useful.
930 \subsubsection pf_storage_examples_modelling Modelling different situations
932 The storage functionality of SimGrid is type-agnostic, that is, the implementation
933 does not presume any type of storage, such as HDDs/SSDs, RAM,
934 CD/DVD devices, USB sticks etc.
936 This allows the user to apply the simulator for a wide variety of scenarios; one
937 common scenario would be the access of remote RAM.
939 #### Modelling the access of remote RAM ####
941 How can this be achieved in SimGrid? Let's assume we have a setup where three hosts
942 (HostA, HostB, HostC) need to access remote RAM:
952 An easy way to model this scenario is to setup and define the RAM via the
953 \ref pf_storage_entity_storage "storage" and \ref pf_storage_entity_storage_type "storage type"
954 entities and attach it to a remote dummy host; then, every host can have their own links
955 to this host (modelling for instance certain scenarios, such as PCIe ...)
960 RAM - Dummy -- Host B
965 Now, if read from this storage, the host that mounts this storage
966 communicates to the dummy host which reads from RAM and
967 sends the information back.
970 \section pf_routing Routing
972 To achieve high performance, the routing tables used within SimGrid are
973 static. This means that routing between two nodes is calculated once
974 and will not change during execution. The SimGrid team chose to use this
975 approach as it is rare to have a real deficiency of a resource;
976 most of the time, a communication fails because the links experience too much
977 congestion and hence, your connection stops before the timeout or
978 because the computer designated to be the destination of that message
981 We also chose to use shortest paths algorithms in order to emulate
982 routing. Doing so is consistent with the reality: [RIP](https://en.wikipedia.org/wiki/Routing_Information_Protocol),
983 [OSPF](https://en.wikipedia.org/wiki/Open_Shortest_Path_First), [BGP](https://en.wikipedia.org/wiki/Border_Gateway_Protocol)
984 are all calculating shortest paths. They do require some time to converge, but
985 eventually, when the routing tables have stabilized, your packets will follow
988 \subsection pf_rm Routing models
990 For each AS, you must define explicitly which routing model will
991 be used. There are 3 different categories for routing models:
993 1. \ref pf_routing_model_shortest_path "Shortest-path" based models: SimGrid calculates shortest
994 paths and manages them. Behaves more or less like most real life
996 2. \ref pf_routing_model_manual "Manually-entered" route models: you have to define all routes
997 manually in the platform description file; this can become
998 tedious very quickly, as it is very verbose.
999 Consistent with some manually managed real life routing.
1000 3. \ref pf_routing_model_simple "Simple/fast models": those models offer fast, low memory routing
1001 algorithms. You should consider to use this type of model if
1002 you can make some assumptions about your AS.
1003 Routing in this case is more or less ignored.
1005 \subsubsection pf_raf The router affair
1007 Using routers becomes mandatory when using shortest-path based
1008 models or when using the bindings to the ns-3 packet-level
1009 simulator instead of the native analytical network model implemented
1012 For graph-based shortest path algorithms, routers are mandatory, because these
1013 algorithms require a graph as input and so we need to have source and
1014 destination for each edge.
1016 Routers are naturally an important concept ns-3 since the
1017 way routers run the packet routing algorithms is actually simulated.
1018 SimGrid's analytical models however simply aggregate the routing time
1019 with the transfer time.
1021 So why did we incorporate routers in SimGrid? Rebuilding a graph representation
1022 only from the route information turns out to be a very difficult task, because
1023 of the missing information about how routes intersect. That is why we
1024 introduced routers, which are simply used to express these intersection points.
1025 It is important to understand that routers are only used to provide topological
1028 To express this topological information, a <b>route</b> has to be
1029 defined in order to declare which link is connected to a router.
1032 \subsubsection pf_routing_model_shortest_path Shortest-path based models
1034 The following table shows all the models that compute routes using
1035 shortest-paths algorithms are currently available in SimGrid. More detail on how
1036 to choose the best routing model is given in the Section called \"\ref pf_routing_howto_choose_wisely\".
1038 | Name | Description |
1039 | --------------------------------------------------- | -------------------------------------------------------------------------- |
1040 | \ref pf_routing_model_floyd "Floyd" | Floyd routing data. Pre-calculates all routes once |
1041 | \ref pf_routing_model_dijkstra "Dijkstra" | Dijkstra routing data. Calculates routes only when needed |
1042 | \ref pf_routing_model_dijkstracache "DijkstraCache" | Dijkstra routing data. Handles some cache for already calculated routes. |
1044 All those shortest-path models are instanciated in the same way and are
1045 completely interchangeable. Here are some examples:
1047 \anchor pf_routing_model_floyd
1052 <AS id="AS0" routing="Floyd">
1054 <cluster id="my_cluster_1" prefix="c-" suffix=""
1055 radical="0-1" power="1000000000" bw="125000000" lat="5E-5"
1056 router_id="router1"/>
1058 <AS id="AS1" routing="None">
1059 <host id="host1" power="1000000000"/>
1062 <link id="link1" bandwidth="100000" latency="0.01"/>
1064 <ASroute src="my_cluster_1" dst="AS1"
1067 <link_ctn id="link1"/>
1073 ASroute given at the end gives a topological information: link1 is
1074 between router1 and host1.
1076 #### Example platform files ####
1078 This is an automatically generated list of example files that use the Floyd
1079 routing model (the path is given relative to SimGrid's source directory)
1081 \verbinclude example_filelist_routing_floyd
1083 \anchor pf_routing_model_dijkstra
1086 #### Example platform files ####
1088 This is an automatically generated list of example files that use the Dijkstra
1089 routing model (the path is given relative to SimGrid's source directory)
1091 \verbinclude example_filelist_routing_dijkstra
1095 <AS id="AS_2" routing="Dijkstra">
1096 <host id="AS_2_host1" power="1000000000"/>
1097 <host id="AS_2_host2" power="1000000000"/>
1098 <host id="AS_2_host3" power="1000000000"/>
1099 <link id="AS_2_link1" bandwidth="1250000000" latency="5E-4"/>
1100 <link id="AS_2_link2" bandwidth="1250000000" latency="5E-4"/>
1101 <link id="AS_2_link3" bandwidth="1250000000" latency="5E-4"/>
1102 <link id="AS_2_link4" bandwidth="1250000000" latency="5E-4"/>
1103 <router id="central_router"/>
1104 <router id="AS_2_gateway"/>
1105 <!-- routes providing topological information -->
1106 <route src="central_router" dst="AS_2_host1"><link_ctn id="AS_2_link1"/></route>
1107 <route src="central_router" dst="AS_2_host2"><link_ctn id="AS_2_link2"/></route>
1108 <route src="central_router" dst="AS_2_host3"><link_ctn id="AS_2_link3"/></route>
1109 <route src="central_router" dst="AS_2_gateway"><link_ctn id="AS_2_link4"/></route>
1113 \anchor pf_routing_model_dijkstracache
1114 ### DijkstraCache ###
1116 DijkstraCache example:
1118 <AS id="AS_2" routing="DijkstraCache">
1119 <host id="AS_2_host1" power="1000000000"/>
1121 (platform unchanged compared to upper example)
1124 #### Example platform files ####
1126 This is an automatically generated list of example files that use the DijkstraCache
1127 routing model (the path is given relative to SimGrid's source directory):
1129 Editor's note: At the time of writing, no platform file used this routing model - so
1130 if there are no example files listed here, this is likely to be correct.
1132 \verbinclude example_filelist_routing_dijkstra_cache
1134 \subsubsection pf_routing_model_manual Manually-entered route models
1136 | Name | Description |
1137 | ---------------------------------- | ------------------------------------------------------------------------------ |
1138 | \ref pf_routing_model_full "Full" | You have to enter all necessary routers manually; that is, every single route. This may consume a lot of memory when the XML is parsed and might be tedious to write; i.e., this is only recommended (if at all) for small platforms. |
1140 \anchor pf_routing_model_full
1145 <AS id="AS0" routing="Full">
1146 <host id="host1" power="1000000000"/>
1147 <host id="host2" power="1000000000"/>
1148 <link id="link1" bandwidth="125000000" latency="0.000100"/>
1149 <route src="host1" dst="host2"><link_ctn id="link1"/></route>
1153 #### Example platform files ####
1155 This is an automatically generated list of example files that use the Full
1156 routing model (the path is given relative to SimGrid's source directory):
1158 \verbinclude example_filelist_routing_full
1160 \subsubsection pf_routing_model_simple Simple/fast models
1162 | Name | Description |
1163 | ---------------------------------------- | ------------------------------------------------------------------------------ |
1164 | \ref pf_routing_model_cluster "Cluster" | This is specific to the \ref pf_cluster "<cluster/>" tag and should not be used by the user, as several assumptions are made. |
1165 | \ref pf_routing_model_none "None" | No routing at all. Unless you know what you're doing, avoid using this mode in combination with a non-constant network model. |
1166 | \ref pf_routing_model_vivaldi "Vivaldi" | Perfect when you want to use coordinates. Also see the corresponding \ref pf_P2P_tags "P2P section" below. |
1168 \anchor pf_routing_model_cluster
1172 In this mode, the \ref pf_cabinet "<cabinet/>" tag is available.
1174 #### Example platform files ####
1176 This is an automatically generated list of example files that use the Cluster
1177 routing model (the path is given relative to SimGrid's source directory):
1179 \verbinclude example_filelist_routing_cluster
1181 \anchor pf_routing_model_none
1184 This model does exactly what it's name advertises: Nothing. There is no routing
1185 available within this model and if you try to communicate within the AS that
1186 uses this model, SimGrid will fail unless you have explicitly activated the
1187 \ref options_model_select_network_constant "Constant Network Model" (this model charges
1188 the same for every single communication). It should
1189 be noted, however, that you can still attach an \ref pf_routing_tag_asroute "ASroute",
1190 as is demonstrated in the example below:
1192 \verbinclude platforms/cluster_and_one_host.xml
1194 #### Example platform files ####
1196 This is an automatically generated list of example files that use the None
1197 routing model (the path is given relative to SimGrid's source directory):
1199 \verbinclude example_filelist_routing_none
1202 \anchor pf_routing_model_vivaldi
1205 For more information on how to use the [Vivaldi Coordinates](https://en.wikipedia.org/wiki/Vivaldi_coordinates),
1206 see also Section \ref pf_P2P_tags "P2P tags".
1208 For documentation on how to activate this model (as some initialization must be done
1209 in the simulator), see Section \ref options_model_network_coord "Activating Coordinate Based Routing".
1211 Note that it is possible to combine the Vivaldi routing model with other routing models;
1212 an example can be found in the file \c examples/platforms/cloud.xml. This
1213 examples models an AS using Vivaldi that contains other ASes that use different
1216 #### Example platform files ####
1218 This is an automatically generated list of example files that use the None
1219 routing model (the path is given relative to SimGrid's source directory):
1221 \verbinclude example_filelist_routing_vivaldi
1224 \subsection ps_dec Defining routes
1226 There are currently four different ways to define routes:
1228 | Name | Description |
1229 | ------------------------------------------------- | ----------------------------------------------------------------------------------- |
1230 | \ref pf_routing_tag_route "route" | Used to define route between host/router |
1231 | \ref pf_routing_tag_asroute "ASroute" | Used to define route between different AS |
1232 | \ref pf_routing_tag_bypassroute "bypassRoute" | Used to supersede normal routes as calculated by the network model between host/router; e.g., can be used to use a route that is not the shortest path for any of the shortest-path routing models. |
1233 | \ref pf_routing_tag_bypassasroute "bypassASroute" | Used in the same way as bypassRoute, but for AS |
1235 Basically all those tags will contain an (ordered) list of references
1236 to link that compose the route you want to define.
1238 Consider the example below:
1241 <route src="Alice" dst="Bob">
1242 <link_ctn id="link1"/>
1243 <link_ctn id="link2"/>
1244 <link_ctn id="link3"/>
1248 The route here from host Alice to Bob will be first link1, then link2,
1249 and finally link3. What about the reverse route? \ref pf_routing_tag_route "Route" and
1250 \ref pf_routing_tag_asroute "ASroute" have an optional attribute \c symmetrical, that can
1251 be either \c YES or \c NO. \c YES means that the reverse route is the same
1252 route in the inverse order, and is set to \c YES by default. Note that
1253 this is not the case for bypass*Route, as it is more probable that you
1254 want to bypass only one default route.
1256 For an \ref pf_routing_tag_asroute "ASroute", things are just slightly more complicated, as you have
1257 to give the id of the gateway which is inside the AS you want to access ...
1258 So it looks like this:
1261 <ASroute src="AS1" dst="AS2"
1262 gw_src="router1" gw_dst="router2">
1263 <link_ctn id="link1"/>
1267 gw == gateway, so when any message are trying to go from AS1 to AS2,
1268 it means that it must pass through router1 to get out of the AS, then
1269 pass through link1, and get into AS2 by being received by router2.
1270 router1 must belong to AS1 and router2 must belong to AS2.
1272 \subsubsection pf_linkctn <link_ctn/>
1274 This entity has only one purpose: Refer to an already existing
1275 \ref pf_link "<link/>" when defining a route, i.e., it
1276 can only occur as a child of \ref pf_routing_tag_route "<route/>"
1278 | Attribute name | Mandatory | Values | Description |
1279 | --------------- | --------- | ------ | ----------- |
1280 | id | yes | String | The identifier of the link that should be added to the route. |
1281 | direction | maybe | UP\|DOWN | If the link referenced by \c id has been declared as \ref pf_sharing_policy_fullduplex "FULLDUPLEX", this indicates which direction the route traverses through this link: UP or DOWN. If you don't use FULLDUPLEX, do not use this attribute or SimGrid will not find the right link.
1283 #### Example Files ####
1285 This is an automatically generated list of example files that use the \c <link_ctn/>
1286 entity (the path is given relative to SimGrid's source directory):
1288 \verbinclude example_filelist_xmltag_linkctn
1290 \subsubsection pf_routing_tag_asroute ASroute
1292 The purpose of this entity is to define a route between two ASes.
1293 This is mainly useful when you're in the \ref pf_routing_model_full "Full routing model".
1295 #### Attributes ####
1297 | Attribute name | Mandatory | Values | Description |
1298 | --------------- | --------- | ------ | ----------- |
1299 | src | yes | String | The identifier of the source AS |
1300 | dst | yes | String | See the \c src attribute |
1301 | gw_src | yes | String | The gateway that will be used within the src AS; this can be any \ref pf_host "Host" or \ref pf_router "Router" defined within the src AS. |
1302 | gw_dst | yes | String | Same as \c gw_src, but with the dst AS instead. |
1303 | symmetrical | no | YES\|NO (Default: YES) | If this route is symmetric, the opposite route (from dst to src) will also be declared implicitly. |
1308 <AS id="AS0" routing="Full">
1309 <cluster id="my_cluster_1" prefix="c-" suffix=".me"
1310 radical="0-149" power="1000000000" bw="125000000" lat="5E-5"
1311 bb_bw="2250000000" bb_lat="5E-4"/>
1313 <cluster id="my_cluster_2" prefix="c-" suffix=".me"
1314 radical="150-299" power="1000000000" bw="125000000" lat="5E-5"
1315 bb_bw="2250000000" bb_lat="5E-4"/>
1317 <link id="backbone" bandwidth="1250000000" latency="5E-4"/>
1319 <ASroute src="my_cluster_1" dst="my_cluster_2"
1320 gw_src="c-my_cluster_1_router.me"
1321 gw_dst="c-my_cluster_2_router.me">
1322 <link_ctn id="backbone"/>
1324 <ASroute src="my_cluster_2" dst="my_cluster_1"
1325 gw_src="c-my_cluster_2_router.me"
1326 gw_dst="c-my_cluster_1_router.me">
1327 <link_ctn id="backbone"/>
1332 \subsubsection pf_routing_tag_route route
1334 The principle is the same as for
1335 \ref pf_routing_tag_asroute "ASroute": The route contains a list of links that
1336 provide a path from \c src to \c dst. Here, \c src and \c dst can both be either a
1337 \ref pf_host "host" or \ref pf_router "router". This is mostly useful for the
1338 \ref pf_routing_model_full "Full routing model" as well as for the
1339 \ref pf_routing_model_shortest_path "shortest-paths" based models (as they require
1340 topological information).
1343 | Attribute name | Mandatory | Values | Description |
1344 | --------------- | --------- | ---------------------- | ----------- |
1345 | src | yes | String | The value given to the source's "id" attribute |
1346 | dst | yes | String | The value given to the destination's "id" attribute. |
1347 | symmetrical | no | YES\| NO (Default: YES) | If this route is symmetric, the opposite route (from dst to src) will also be declared implicitly. |
1352 A route in the \ref pf_routing_model_full "Full routing model" could look like this:
1354 <route src="Tremblay" dst="Bourassa">
1355 <link_ctn id="4"/><link_ctn id="3"/><link_ctn id="2"/><link_ctn id="0"/><link_ctn id="1"/><link_ctn id="6"/><link_ctn id="7"/>
1359 A route in the \ref pf_routing_model_shortest_path "Shortest-Path routing model" could look like this:
1361 <route src="Tremblay" dst="Bourassa">
1366 You must only have one link in your routes when you're using them to provide
1367 topological information, as the routes here are simply the edges of the
1368 (network-)graph and the employed algorithms need to know which edge connects
1369 which pair of entities.
1371 \subsubsection pf_routing_tag_bypassasroute bypassASroute
1373 %As said before, once you choose
1374 a model, it (most likely; the constant network model, for example, doesn't) calculates routes for you. But maybe you want to
1375 define some of your routes, which will be specific. You may also want
1376 to bypass some routes defined in lower level AS at an upper stage:
1377 <b>bypassASroute</b> is the tag you're looking for. It allows to
1378 bypass routes defined between already defined between AS (if you want
1379 to bypass route for a specific host, you should just use byPassRoute).
1380 The principle is the same as ASroute : <b>bypassASroute</b> contains
1381 list of links that are in the path between src and dst.
1383 #### Attributes ####
1385 | Attribute name | Mandatory | Values | Description |
1386 | --------------- | --------- | ---------------------- | ----------- |
1387 | src | yes | String | The value given to the source AS's "id" attribute |
1388 | dst | yes | String | The value given to the destination AS's "id" attribute. |
1389 | gw_src | yes | String | The value given to the source gateway's "id" attribute; this can be any host or router within the src AS |
1390 | gw_dst | yes | String | The value given to the destination gateway's "id" attribute; this can be any host or router within the dst AS|
1391 | symmetrical | no | YES\| NO (Default: YES) | If this route is symmetric, the opposite route (from dst to src) will also be declared implicitly. |
1396 <bypassASRoute src="my_cluster_1" dst="my_cluster_2"
1397 gw_src="my_cluster_1_router"
1398 gw_dst="my_cluster_2_router">
1399 <link_ctn id="link_tmp"/>
1403 This example shows that link \c link_tmp (definition not displayed here) directly
1404 connects the router \c my_cluster_1_router in the source cluster to the router
1405 \c my_cluster_2_router in the destination router. Additionally, as the \c symmetrical
1406 attribute was not given, this route is presumed to be symmetrical.
1408 \subsubsection pf_routing_tag_bypassroute bypassRoute
1410 %As said before, once you choose
1411 a model, it (most likely; the constant network model, for example, doesn't) calculates routes for you. But maybe you want to
1412 define some of your routes, which will be specific. You may also want
1413 to bypass some routes defined in lower level AS at an upper stage :
1414 <b>bypassRoute</b> is the tag you're looking for. It allows to bypass
1415 routes defined between <b>host/router</b>. The principle is the same
1416 as route : <b>bypassRoute</b> contains list of links references of
1417 links that are in the path between src and dst.
1419 #### Attributes ####
1421 | Attribute name | Mandatory | Values | Description |
1422 | --------------- | --------- | ---------------------- | ----------- |
1423 | src | yes | String | The value given to the source AS's "id" attribute |
1424 | dst | yes | String | The value given to the destination AS's "id" attribute. |
1425 | symmetrical | no | YES \| NO (Default: YES) | If this route is symmetric, the opposite route (from dst to src) will also be declared implicitly. |
1430 <bypassRoute src="host_1" dst="host_2">
1431 <link_ctn id="link_tmp"/>
1435 This example shows that link \c link_tmp (definition not displayed here) directly
1436 connects host \c host_1 to host \c host_2. Additionally, as the \c symmetrical
1437 attribute was not given, this route is presumed to be symmetrical.
1439 \subsection pb_baroex Basic Routing Example
1441 Let's say you have an AS named AS_Big that contains two other AS, AS_1
1442 and AS_2. If you want to make a host (h1) from AS_1 with another one
1443 (h2) from AS_2 then you'll have to proceed as follows:
1444 \li First, you have to ensure that a route is defined from h1 to the
1445 AS_1's exit gateway and from h2 to AS_2's exit gateway.
1446 \li Then, you'll have to define a route between AS_1 to AS_2. %As those
1447 AS are both resources belonging to AS_Big, then it has to be done
1448 at AS_big level. To define such a route, you have to give the
1449 source AS (AS_1), the destination AS (AS_2), and their respective
1450 gateway (as the route is effectively defined between those two
1451 entry/exit points). Elements of this route can only be elements
1452 belonging to AS_Big, so links and routers in this route should be
1453 defined inside AS_Big. If you choose some shortest-path model,
1454 this route will be computed automatically.
1456 %As said before, there are mainly 2 tags for routing :
1457 \li <b>ASroute</b>: to define routes between two <b>AS</b>
1458 \li <b>route</b>: to define routes between two <b>host/router</b>
1460 %As we are dealing with routes between AS, it means that those we'll
1461 have some definition at AS_Big level. Let consider AS_1 contains 1
1462 host, 1 link and one router and AS_2 3 hosts, 4 links and one router.
1463 There will be a central router, and a cross-like topology. At the end
1464 of the crosses arms, you'll find the 3 hosts and the router that will
1465 act as a gateway. We have to define routes inside those two AS. Let
1466 say that AS_1 contains full routes, and AS_2 contains some Floyd
1467 routing (as we don't want to bother with defining all routes). %As
1468 we're using some shortest path algorithms to route into AS_2, we'll
1469 then have to define some <b>route</b> to gives some topological
1470 information to SimGrid. Here is a file doing it all :
1473 <AS id="AS_Big" routing="Dijkstra">
1474 <AS id="AS_1" routing="Full">
1475 <host id="AS_1_host1" power="1000000000"/>
1476 <link id="AS_1_link" bandwidth="1250000000" latency="5E-4"/>
1477 <router id="AS_1_gateway"/>
1478 <route src="AS_1_host1" dst="AS_1_gateway">
1479 <link_ctn id="AS_1_link"/>
1482 <AS id="AS_2" routing="Floyd">
1483 <host id="AS_2_host1" power="1000000000"/>
1484 <host id="AS_2_host2" power="1000000000"/>
1485 <host id="AS_2_host3" power="1000000000"/>
1486 <link id="AS_2_link1" bandwidth="1250000000" latency="5E-4"/>
1487 <link id="AS_2_link2" bandwidth="1250000000" latency="5E-4"/>
1488 <link id="AS_2_link3" bandwidth="1250000000" latency="5E-4"/>
1489 <link id="AS_2_link4" bandwidth="1250000000" latency="5E-4"/>
1490 <router id="central_router"/>
1491 <router id="AS_2_gateway"/>
1492 <!-- routes providing topological information -->
1493 <route src="central_router" dst="AS_2_host1"><link_ctn id="AS_2_link1"/></route>
1494 <route src="central_router" dst="AS_2_host2"><link_ctn id="AS_2_link2"/></route>
1495 <route src="central_router" dst="AS_2_host3"><link_ctn id="AS_2_link3"/></route>
1496 <route src="central_router" dst="AS_2_gateway"><link_ctn id="AS_2_link4"/></route>
1498 <link id="backbone" bandwidth="1250000000" latency="5E-4"/>
1500 <ASroute src="AS_1" dst="AS_2"
1501 gw_src="AS_1_gateway"
1502 gw_dst="AS_2_gateway">
1503 <link_ctn id="backbone"/>
1508 \section pf_other_tags Tags not (directly) describing the platform
1510 The following tags can be used inside a \<platform\> tag even if they are not
1511 directly describing the platform:
1512 \li \ref pf_config "config": it allows you to pass some configuration stuff like, for
1513 example, the network model and so on. It follows the
1514 \li \ref pf_include "include": allows you to include another file into the current one.
1516 \subsection pf_config config
1518 The only purpose of this tag is to contain the \c prop tags, as described below.
1519 These tags will then configure the options as described by Section \ref options.
1522 #### Attributes ####
1524 | Attribute name | Mandatory | Values | Description |
1525 | --------------- | --------- | ---------------------- | ----------- |
1526 | id | yes | String | The identifier of the config tag when referring to id; this is basically useless, though. |
1528 #### Possible children ####
1530 Tag name | Description | Documentation
1531 ------------ | ----------- | -------------
1532 \<prop/\> | The prop tag allows you to define different configuration options following the attribute/value schema. See the \ref options page. | N/A
1537 <?xml version='1.0'?>
1538 <!DOCTYPE platform SYSTEM "http://simgrid.gforge.inria.fr/simgrid.dtd">
1539 <platform version="4">
1540 <config id="General">
1541 <prop id="maxmin/precision" value="0.000010"></prop>
1542 <prop id="cpu/optim" value="TI"></prop>
1543 <prop id="host/model" value="compound"></prop>
1544 <prop id="network/model" value="SMPI"></prop>
1545 <prop id="path" value="~/"></prop>
1546 <prop id="smpi/bw-factor" value="65472:0.940694;15424:0.697866;9376:0.58729"></prop>
1549 <AS id="AS0" routing="Full">
1553 \subsection pf_include include
1555 Even if it can be used in other contexts, this tag was originally created
1556 to be used with \ref pf_trace. The idea was to have a file describing the
1557 platform, and another file attaching traces of a given period to the platform.
1559 The drawback is that the file chuncks that will be included do not
1560 constitute valid XML files. This may explain why this feature was never really
1561 used in practice (as far as we know). Other mechanisms, such as the ability to load
1562 several platform files one after the other, could be considered in the future.
1564 In the meanwhile, the \c include tag allows you to import other platforms into your
1565 local file. This is done with the intention to help people
1566 combine their different AS and provide new platforms. Those files
1567 should contain XML that consists of
1568 \ref pf_include "include", \ref pf_cluster "cluster", \ref pf_peer "peer", \ref pf_As "AS", \ref pf_trace "trace", \ref pf_trace "tags".
1570 Do not forget to close the tag to make it work, or you will end up with an invalid XML file.
1572 #### Attributes ####
1574 | Attribute name | Mandatory | Values | Description |
1575 | --------------- | --------- | ---------------------- | ----------- |
1576 | file | yes | String | Filename of the path you want to include with either relative or absolute path. |
1581 The following example includes two files, clusterA.xml and clusterB.xml and
1582 combines them two one platform file; all hosts, routers etc. defined in
1583 each of them will then be usable.
1586 <?xml version='1.0'?>
1587 <!DOCTYPE platform SYSTEM "http://simgrid.gforge.inria.fr/simgrid/simgrid.dtd">
1588 <platform version="4">
1589 <AS id="main" routing="Full">
1590 <include file="clusterA.xml"></include>
1591 <include file="clusterB.xml"></include>
1596 \subsection pf_trace trace and trace_connect
1598 Both tags are an alternate way to pass files containing information on
1599 availability, state etc. to an entity. (See also, for instance, Section \ref
1600 pf_host_churn "Churn", as described for the host entity.) Instead of referring
1601 to the file directly in the host, link, or cluster tag, you proceed by defining
1602 a trace with an id corresponding to a file, later a host/link/cluster, and
1603 finally using trace_connect you say that the file trace must be used by the
1610 <AS id="AS0" routing="Full">
1611 <host id="bob" power="1000000000"/>
1613 <trace id="myTrace" file="bob.trace" periodicity="1.0"/>
1614 <trace_connect trace="myTrace" element="bob" kind="POWER"/>
1618 The order here is important. \c trace_connect must come
1619 after the elements \c trace and \c host, as both the host
1620 and the trace definition must be known when \c trace_connect
1621 is parsed; the order of \c trace and \c host is arbitrary.
1624 #### \c trace attributes ####
1627 | Attribute name | Mandatory | Values | Description |
1628 | --------------- | --------- | ---------------------- | ----------- |
1629 | id | yes | String | Identifier of this trace; this is the name you pass on to \c trace_connect. |
1630 | file | no | String | Filename of the file that contains the information - the path must follow the style of your OS. You can omit this, but then you must specifiy the values inside of <trace> and </trace> - see the example below. |
1631 | trace_periodicity | yes | String | This is the same as for \ref pf_host "hosts" (see there for details) |
1633 Here is an example of trace when no file name is provided:
1636 <trace id="myTrace" periodicity="1.0">
1643 #### \c trace_connect attributes ####
1645 | Attribute name | Mandatory | Values | Description |
1646 | --------------- | --------- | ---------------------- | ----------- |
1647 | kind | no | HOST_AVAIL\|POWER\|<br/>LINK_AVAIL\|BANDWIDTH\|LATENCY (Default: HOST_AVAIL) | Describes the kind of trace. |
1648 | trace | yes | String | Identifier of the referenced trace (specified of the trace's \c id attribute) |
1649 | element | yes | String | The identifier of the referenced entity as given by its \c id attribute |
1651 \section pf_hints Hints, tips and frequently requested features
1653 Now you should know at least the syntax and be able to create a
1654 platform by your own. However, after having ourselves wrote some platforms, there
1655 are some best practices you should pay attention to in order to
1656 produce good platform and some choices you can make in order to have
1657 faster simulations. Here's some hints and tips, then.
1659 @subsection Finding the platform example that you need
1661 Most platform files that we ship are in the @c examples/platforms
1662 folder. The good old @c grep tool can find the examples you need when
1663 wondering on a specific XML tag. Here is an example session searching
1664 for @ref pf_trace "trace_connect":
1667 % cd examples/platforms
1668 % grep -R -i -n --include="*.xml" "trace_connect" .
1669 ./two_hosts_platform_with_availability_included.xml:26:<trace_connect kind="SPEED" trace="A" element="Cpu A"/>
1670 ./two_hosts_platform_with_availability_included.xml:27:<trace_connect kind="HOST_AVAIL" trace="A_failure" element="Cpu A"/>
1671 ./two_hosts_platform_with_availability_included.xml:28:<trace_connect kind="SPEED" trace="B" element="Cpu B"/>
1672 ./two_hosts.xml:17: <trace_connect trace="Tremblay_power" element="Tremblay" kind="SPEED"/>
1675 \subsection pf_as_h AS Hierarchy
1676 The AS design allows SimGrid to go fast, because computing route is
1677 done only for the set of resources defined in this AS. If you're using
1678 only a big AS containing all resource with no AS into it and you're
1679 using Full model, then ... you'll loose all interest into it. On the
1680 other hand, designing a binary tree of AS with, at the lower level,
1681 only one host, then you'll also loose all the good AS hierarchy can
1682 give you. Remind you should always be "reasonable" in your platform
1683 definition when choosing the hierarchy. A good choice if you try to
1684 describe a real life platform is to follow the AS described in
1685 reality, since this kind of trade-off works well for real life
1688 \subsection pf_exit_as Exit AS: why and how
1689 Users that have looked at some of our platforms may have notice a
1690 non-intuitive schema ... Something like that :
1694 <AS id="AS_4" routing="Full">
1695 <AS id="exitAS_4" routing="Full">
1696 <router id="router_4"/>
1698 <cluster id="cl_4_1" prefix="c_4_1-" suffix="" radical="1-20" power="1000000000" bw="125000000" lat="5E-5" bb_bw="2250000000" bb_lat="5E-4"/>
1699 <cluster id="cl_4_2" prefix="c_4_2-" suffix="" radical="1-20" power="1000000000" bw="125000000" lat="5E-5" bb_bw="2250000000" bb_lat="5E-4"/>
1700 <link id="4_1" bandwidth="2250000000" latency="5E-5"/>
1701 <link id="4_2" bandwidth="2250000000" latency="5E-5"/>
1702 <link id="bb_4" bandwidth="2250000000" latency="5E-4"/>
1703 <ASroute src="cl_4_1"
1705 gw_src="c_4_1-cl_4_1_router"
1706 gw_dst="c_4_2-cl_4_2_router">
1707 <link_ctn id="4_1"/>
1708 <link_ctn id="bb_4"/>
1709 <link_ctn id="4_2"/>
1711 <ASroute src="cl_4_1"
1713 gw_src="c_4_1-cl_4_1_router"
1715 <link_ctn id="4_1"/>
1716 <link_ctn id="bb_4"/>
1718 <ASroute src="cl_4_2"
1720 gw_src="c_4_2-cl_4_2_router"
1722 <link_ctn id="4_2"/>
1723 <link_ctn id="bb_4"/>
1728 In the AS_4, you have an exitAS_4 defined, containing only one router,
1729 and routes defined to that AS from all other AS (as cluster is only a
1730 shortcut for an AS, see cluster description for details). If there was
1731 an upper AS, it would define routes to and from AS_4 with the gateway
1732 router_4. It's just because, as we did not allowed (for performances
1733 issues) to have routes from an AS to a single host/router, you have to
1734 enclose your gateway, when you have AS included in your AS, within an
1735 AS to define routes to it.
1737 \subsection pf_P2P_tags P2P or how to use coordinates
1738 SimGrid allows you to use some coordinated-based system, like vivaldi,
1739 to describe a platform. The main concept is that you have some peers
1740 that are located somewhere: this is the function of the
1741 <b>coordinates</b> of the \<peer\> or \<host\> tag. There's nothing
1742 complicated in using it, here is an example of it:
1745 <?xml version='1.0'?>
1746 <!DOCTYPE platform SYSTEM "http://simgrid.gforge.inria.fr/simgrid.dtd">
1747 <platform version="4">
1749 <config id="General">
1750 <prop id="network/coordinates" value="yes"></prop>
1752 <AS id="AS0" routing="Vivaldi">
1753 <host id="100030591" coordinates="25.5 9.4 1.4" power="1.5Gf" />
1754 <host id="100036570" coordinates="-12.7 -9.9 2.1" power="7.3Gf" />
1756 <host id="100429957" coordinates="17.5 6.7 18.8" power="8.3Gf" />
1761 Coordinates are then used to calculate latency (in microseconds)
1762 between two hosts by calculating the distance between the two hosts
1763 coordinates with the following formula: distance( (x1, y1, z1), (x2,
1764 y2, z2) ) = euclidian( (x1,y1), (x2,y2) ) + abs(z1) + abs(z2)
1766 In other words, we take the euclidian distance on the two first
1767 dimensions, and then add the absolute values found on the third
1768 dimension. This may seem strange, but it was found to allow better
1769 approximations of the latency matrices (see the paper describing
1772 Note that the previous example defines a routing directly between hosts but it could be also used to define a routing between AS.
1773 That is for example what is commonly done when using peers (see Section \ref pf_peer).
1775 <?xml version='1.0'?>
1776 <!DOCTYPE platform SYSTEM "http://simgrid.gforge.inria.fr/simgrid.dtd">
1777 <platform version="4">
1779 <config id="General">
1780 <prop id="network/coordinates" value="yes"></prop>
1782 <AS id="AS0" routing="Vivaldi">
1783 <peer id="peer-0" coordinates="173.0 96.8 0.1" power="730Mf" bw_in="13.38MBps" bw_out="1.024MBps" lat="500us"/>
1784 <peer id="peer-1" coordinates="247.0 57.3 0.6" power="730Mf" bw_in="13.38MBps" bw_out="1.024MBps" lat="500us" />
1785 <peer id="peer-2" coordinates="243.4 58.8 1.4" power="730Mf" bw_in="13.38MBps" bw_out="1.024MBps" lat="500us" />
1789 In such a case though, we connect the AS created by the <b>peer</b> tag with the Vivaldi routing mechanism.
1790 This means that to route between AS1 and AS2, it will use the coordinates of router_AS1 and router_AS2.
1791 This is currently a convention and we may offer to change this convention in the DTD later if needed.
1792 You may have noted that conveniently, a peer named FOO defines an AS named FOO and a router named router_FOO, which is why it works seamlessly with the <b>peer</b> tag.
1795 \subsection pf_routing_howto_choose_wisely Choosing wisely the routing model to use
1798 Choosing wisely the routing model to use can significantly fasten your
1799 simulation/save your time when writing the platform/save tremendous
1800 disk space. Here is the list of available model and their
1801 characteristics (lookup : time to resolve a route):
1803 \li <b>Full</b>: Full routing data (fast, large memory requirements,
1805 \li <b>Floyd</b>: Floyd routing data (slow initialization, fast
1806 lookup, lesser memory requirements, shortest path routing only).
1807 Calculates all routes at once at the beginning.
1808 \li <b>Dijkstra</b>: Dijkstra routing data (fast initialization, slow
1809 lookup, small memory requirements, shortest path routing only).
1810 Calculates a route when necessary.
1811 \li <b>DijkstraCache</b>: Dijkstra routing data (fast initialization,
1812 fast lookup, small memory requirements, shortest path routing
1813 only). Same as Dijkstra, except it handles a cache for latest used
1815 \li <b>None</b>: No routing (usable with Constant network only).
1816 Defines that there is no routes, so if you try to determine a
1817 route without constant network within this AS, SimGrid will raise
1819 \li <b>Vivaldi</b>: Vivaldi routing, so when you want to use coordinates
1820 \li <b>Cluster</b>: Cluster routing, specific to cluster tag, should
1823 \subsection pf_switch I want to describe a switch but there is no switch tag!
1825 Actually we did not include switch tag. But when you're trying to
1826 simulate a switch, assuming
1827 fluid bandwidth models are used (which SimGrid uses by default unless
1828 ns-3 or constant network models are activated), the limiting factor is
1829 switch backplane bandwidth. So, essentially, at least from
1830 the simulation perspective, a switch is similar to a
1831 link: some device that is traversed by flows and with some latency and
1832 so,e maximum bandwidth. Thus, you can simply simulate a switch as a
1834 can be connected to this "switch", which is then included in routes just
1838 \subsection pf_multicabinets I want to describe multi-cabinets clusters!
1840 You have several possibilities, as usual when modeling things. If your
1841 cabinets are homogeneous and the intercabinet network negligible for
1842 your study, you should just create a larger cluster with all hosts at
1845 In the rare case where your hosts are not homogeneous between the
1846 cabinets, you can create your cluster completely manually. For that,
1847 create an As using the Cluster routing, and then use one
1848 <cabinet> for each cabinet. This cabinet tag can only be used an
1849 As using the Cluster routing schema, and creating
1851 Be warned that creating a cluster manually from the XML with
1852 <cabinet>, <backbone> and friends is rather tedious. The
1853 easiest way to retrieve some control of your model without diving into
1854 the <cluster> internals is certainly to create one separate
1855 <cluster> per cabinet and interconnect them together. This is
1856 what we did in the G5K example platform for the Graphen cluster.
1858 \subsection pf_platform_multipath I want to express multipath routing in platform files!
1860 It is unfortunately impossible to express the fact that there is more
1861 than one routing path between two given hosts. Let's consider the
1862 following platform file:
1865 <route src="A" dst="B">
1868 <route src="B" dst="C">
1871 <route src="A" dst="C">
1876 Although it is perfectly valid, it does not mean that data traveling
1877 from A to C can either go directly (using link 3) or through B (using
1878 links 1 and 2). It simply means that the routing on the graph is not
1879 trivial, and that data do not following the shortest path in number of
1880 hops on this graph. Another way to say it is that there is no implicit
1881 in these routing descriptions. The system will only use the routes you
1882 declare (such as <route src="A" dst="C"><link_ctn
1883 id="3"/></route>), without trying to build new routes by aggregating
1886 You are also free to declare platform where the routing is not
1887 symmetrical. For example, add the following to the previous file:
1890 <route src="C" dst="A">
1896 This makes sure that data from C to A go through B where data from A
1897 to C go directly. Don't worry about realism of such settings since
1898 we've seen ways more weird situation in real settings (in fact, that's
1899 the realism of very regular platforms which is questionable, but
1900 that's another story).
