RE: [ogsa-wg] Paper proposing "evolutionary vertical design efforts"
Hi Marvin, Thanks for documenting and explaining what you see as the basic HPC case. I agree with your observations and what you have listed are what most, if not all HPC cluster systems I know of, implement. I think there is something else that may be apparent. If I may suggest that the bulleted list with the exception of the "11th (?)" bullet would nicely fit into the BES situation. If I look at the cluster you describe (i.e. the client perspective with queued, running and finished) it seem to me that that would be a BES container (at least from an interface perspective and where the scheduler that you are submitting the jobs to is the abstract representation of the BES "container"). We had discussed this recursive container concept in the EMS discussion in OGSA V1.0. This seems especially true since you advocate that in this simple HPC case a "user" would not have visibility into the granular elements of what makes a cluster and the operations would essentially be within this "black box". Would it make sense to add most of these notions to the BES discussion and the scheduler aspects would fit in RSS (where they could provide a basic RSS interface)? Putting these together with some minor additional specs may be the basic HPC profile. Am I missing something? Thanks! Ravi ________________________________ From: owner-ogsa-wg@ggf.org [mailto:owner-ogsa-wg@ggf.org] On Behalf Of Marvin Theimer Sent: Monday, March 13, 2006 2:42 PM To: Ian Foster; ogsa-wg@ggf.org Cc: Marvin Theimer Subject: RE: [ogsa-wg] Paper proposing "evolutionary vertical design efforts" Hi; Ian, you are correct that I view job submission to a cluster as being one of the simplest, and hence most basic, HPC use cases to start with. Or, to be slightly more general, I view job submission to a "black box" that can run jobs - be it a cluster or an SMP or an SGI NUMA machine or what-have-you - as being the simplest and hence most basic HPC use case to start with. The key distinction for me is that the internals of the "box" are for the most part not visible to the client, at least as far as submitting and running compute jobs is concerned. There may well be a separate interface for dealing with things like system management, but I want to explicitly separate those things out in order to allow for use of "boxes" that might be managed by proprietary means or by means obeying standards that a particular job submission client is unfamiliar with. I think the use case that Ravi Subramaniam posted to this mailing list back on 2/17 is a good one to start a discussion around. However, I'd like to present it from a different point-of-view than he did. The manner in which the use case is currently presented emphasizes all the capabilities and services needed to handle the fully general case of submitting a batch job to a computing utility/service. That's a great way of producing a taxonomy against which any given system or design can be compared to see what it has to offer. I would argue that the next step is to ask what's the simplest subset that represents a useful system/design and how should one categorize the various capabilities and services he has identified so as to arrive at meaningful components that can be selectively used to obtain progressively more capable systems. Another useful exercise to do is to examine existing job scheduling systems in order to understand what they provide. Since in the real world we will have to deal with the legacy of existing systems it will be important to understand how they relate to the use cases we explore. In the same vein, it will be important to take into account and understand other existing infrastructures that people use that are related to HPC use cases. I'm thinking of things like security infrastructures, directory services, and so forth. From the point-of-view of managing complexity and reducing total-cost-of-ownership, it will be important to understand the extent to which existing infrastructure and services can be reused rather than reinvented. To kick off a discussion around the topic of a minimalist HPC use case, I present a straw man description of such below and then present a first attempt at categorizing various areas of extension. The categorization of extension areas is not meant to be complete or even all that carefully thought-out as far as componentization boundaries are concerned; it is merely meant to be a first contribution to get the discussion going. A basic HPC use case: Compute cluster embedded within an organization. * This is your basic batch job scheduling scenario. Only a very basic state transition diagram is visible to the client, with the following states for a job: queued, running, finished. Additional states -- and associated state transition request operations and functionality -- are not supported. Examples of additional states and associated functionality include suspension of jobs and migration of jobs. * Only "standard" resources can be described, for example: number of cpus/nodes needed, memory requirements, disk requirements, etc. (think resources that are describable by JSDL). * Once a job has been submitted it can be cancelled, but its resource requests can't be modified. * A distributed file system is accessible from client desktop machines and client file servers, as well as compute nodes of the compute cluster. This implies that no data staging is required, that programs can be (for the most part) executed from existing file system locations, and that no program "provisioning" is required (since you can execute them from wherever they are already installed). Thus in this use case all data transfer and program installation operations are the responsibility of the user. * Users already have accounts within the existing security infrastructure (e.g. Kerberos). They would like to use these and not have to create/manage additional authentication/authorization credentials (at least at the level that is visible to them). * The job scheduling service resides at a well-known network name and it is aware of the compute cluster and its resources by "private" means (e.g. it runs on the head node of the cluster and employs private means to monitor and control the resources of the cluster). This implies that there is no need for any sort of directory services for finding the compute cluster or the resources it represents other than basic DNS. * Compute cluster system management is opaque to users and is the concern of the compute cluster's owners. This implies that system management is not part of the compute cluster's public job scheduling interface. This also implies that there is no need for a logging interface to the service. I assume that application-level logging can be done by means of libraries that write to client files; i.e. that there is no need for any sort of special system support for logging. * A simple polling-based interface is the simplest form of interface to something like a job scheduling service. However, a simple call-back notification interface is a very useful addition that potentially provides substantial performance benefits since it can enable the avoidance of lots of unnecessary network traffic. Only job state changes result in notification messages. * There are no notions of fault tolerance. Jobs that fail must be resubmitted by the client. Neither the cluster head node nor its compute nodes are fault tolerant. I do expect the client software to return an indication of failure-due-system-fault when appropriate. (Note that this may also occur when things like network partitions occur.) * One does need some notion of how to deal with orphaned resources and jobs. The notion of job lifetime and post-expiration garbage collection is a natural approach here. * The scheduling service provides a fixed set of scheduling policies, with only a few basic choices (or maybe even just one), such as FIFO or round-robin. There is no notion, in general, of SLAs (which are a form of scheduling policy). * Enough information must be returned to the client when a job finishes to enable basic accounting functionality. This means things like total wall-clock time the job ran and a summary of resources used. There is not a need for the interface to support any sort of grouping of accounting information. That is, jobs do not need to be associated with projects, groups, or other accounting entities and the job scheduling service is not responsible for tracking accounting information across such entities. As long as basic resource utilization information is returnable for each job, accounting can be done externally to the job scheduling service. I do assume that jobs can be uniquely identified by some means and can be uniquely associated with some principal entity existing in the overall system, such as a user name. * Just as there is no notion of requiring the job scheduling service to track any but the most basic job-level accounting information, there is no notion of the service enforcing quotas on jobs. * Although it is generally useful to separate the notions of resource reservation from resource usage (e.g. to enable interactive and debugging use of resources), it is not a necessity for the most basic of job scheduling services. * There is no notion of tying multiple jobs together, either to support things like dependency graphs or to support things like workflows. Such capabilities must be implemented by clients of the job scheduling service. Interesting extension areas: * Additional scheduling policies o Weighted fair-share, ... o Multiple queues o SLAs o ... * Extended resource descriptions o Additional resource types, such as GPUs o Additional types of compute resources, such as desktop computers o Condor-style class ads * Extended job descriptions (as returned to requesting clients and sys admins) * Additional classes of security credentials * Reservations separated from execution o Enabling interactive and debugging jobs o Support for multiple competing schedulers (incl. desktop cycle stealing and market-based approaches to scheduling compute resources) * Ability to modify jobs during their existence * Fault tolerance o Automatic rescheduling of jobs that failed due to system faults o Highly available resources: This is partly a policy statement by a scheduling service about its characteristics and partly the ability to rebind clients to migrated service endpoints * Extended state transition diagrams and associated functionalities o Job suspension o Job migration o ... * Accounting & quotas * Operating on arrays of jobs * Meta-schedulers, multiple schedulers, and ecologies and hierarchies of multiple schedulers o Meta-schedulers * Hierarchical job scheduling with a meta-scheduler as the only entry point; forwarding jobs to the meta-scheduler from other subsidiary schedulers o Condor-style matchmaking * Directory services o Using existing directory services o Abstract directory service interface(s) * Data transfer topics o Application data staging * Naming * Efficiency * Convenience * Cleanup o Program staging/provisioning * Description * Installation * Cleanup Marvin. ________________________________ From: Ian Foster [mailto:foster@mcs.anl.gov] Sent: Monday, February 20, 2006 9:20 AM To: Marvin Theimer; ogsa-wg@ggf.org Cc: Marvin Theimer; Savas Parastatidis; Tony Hey; Marty Humphrey; gcf@grids.ucs.indiana.edu Subject: Re: [ogsa-wg] Paper proposing "evolutionary vertical design efforts" Dear All: The most important thing to understand at this point (IMHO) is the scope of this "HPC use case," as this will determine just how minimal we can be. I get the impression that the principal goal may be "job submission to a cluster." Is that correct? How do we start to circumscribe the scope more explicitly? Ian. At 05:45 AM 2/16/2006 -0800, Marvin Theimer wrote: Enclosed is a paper that advocates an additional set of activities that the authors believe that the OGSA working groups should engage in. Broadly speaking, the OGSA and related working groups are already doing a bunch of important things: * There is broad exploration of the big picture, including enumeration of use cases, taxonomy of areas, identification of research issues, etc. * There is work going on in each of the horizontal areas that have been identified, such as EMS, data services, etc. * There is working going around individual specifications, such as BES, JSDL, etc. Given that individual specifications are beginning to come to fruition, the authors believe it is time to also start defining vertical profilesthat precisely describe how groups of individual specifications should be employed to implement specific use cases in an interoperable manner. The authors also believe that the process of defining these profiles offers an opportunity to close the design loopby relating the various on-going protocol and standards efforts back to the use cases in a very concrete manner. This provides an end-to-end setting in which to identify holes and issues that might require additional protocols and/or (incremental) changes to existing protocols. The paper introduces both the general notion of doing focused vertical design effortsand then focuses on a specific vertical design effort, namely a minimal HPC design. The paper derives a specific HPC design in a first principlesmanner since the authors believe that this increases the chances of identifying issues. As a consequence, existing specifications and the activities of existing working groups are not mentioned and this paper is not an attempt to actually define a specifications profile. Also, the absence of references to existing work is not meant to imply that such work is in any way irrelevant or inappropriate. The paper should be viewed as a first abstract attempt to propose a new kind of activity within OGSA. The expectation is that future open discussions and publications will explore the concrete details of such a proposal. This paper was recently sent to a few key individuals in order to get feedback from them before submitting it to the wider GGF community. Unfortunately that process took longer than intended and some members of the community may have already seen a copy of the paper without knowing the context within it was written. This email should hopefully dispel any misconceptions that may have occurred. For those people who will be around on for the F2F meetings on Friday, Marvin Theimer will be giving a talk on the contents of this paper at a time and place to be announced. Marvin Theimer, Savas Parastatidis, Tony Hey, Marty Humphrey, Geoffrey Fox _______________________________________________________________ Ian Foster www.mcs.anl.gov/~foster Math & Computer Science Div. Dept of Computer Science Argonne National Laboratory The University of Chicago Argonne, IL 60439, U.S.A. Chicago, IL 60637, U.S.A. Tel: 630 252 4619 Fax: 630 252 1997 Globus Alliance, www.globus.org <http://www.globus.org/>
Ravi, It may well be that things fit nicely into what BES and RSS are targeting and that would be great. But before we get to that discussion, I'd like to still spend some more time looking at use cases. In particular, there are several goals that I think we need to pursue via an examination of use cases: * Identify the simplest case that we expect everyone to support. The simpler this case is the easier it will be for interested parties - including most importantly existing legacy parties - to participate in an HPC grid environment, at least at some basic level. * Identify common cases. One complaint I've repeatedly heard from various people in the distributed systems and grid communities is that most existing and proposed solutions don't make it easy to implement or employ the common cases. We should be identifying which use cases represent important common cases so that we can then ask whether and how our solutions will make those as easy as possible to do. * Understand how various more complicated use cases relate to each other. I assert that the key to designing a successful HPC grid solution will be to come up with an approach that enables both evolution and selective extension to occur in a straight-forward and well-defined manner. The combination of a simple base case and a solid approach to evolution and extensibility will enable both everyone to participate as well as organic growth of richer functionality among those parties that need it or are interested in it. So, I propose that the next steps we should take is to start exploring various extended HPC use cases with an intent to identifying common cases and relationships among them. My bulleted list of extensions was intended as an initial unprioritized list of extension subjects, derived in part from the compendium of capabilities listed in the use case you originally posted. I will post my own thoughts on common cases and relationships among extensions in a future email. Marvin. ________________________________ From: Subramaniam, Ravi [mailto:ravi.subramaniam@intel.com] Sent: Monday, March 13, 2006 8:40 PM To: Marvin Theimer; Ian Foster; ogsa-wg@ggf.org Subject: RE: [ogsa-wg] Paper proposing "evolutionary vertical design efforts" Hi Marvin, Thanks for documenting and explaining what you see as the basic HPC case. I agree with your observations and what you have listed are what most, if not all HPC cluster systems I know of, implement. I think there is something else that may be apparent. If I may suggest that the bulleted list with the exception of the "11th (?)" bullet would nicely fit into the BES situation. If I look at the cluster you describe (i.e. the client perspective with queued, running and finished) it seem to me that that would be a BES container (at least from an interface perspective and where the scheduler that you are submitting the jobs to is the abstract representation of the BES "container"). We had discussed this recursive container concept in the EMS discussion in OGSA V1.0. This seems especially true since you advocate that in this simple HPC case a "user" would not have visibility into the granular elements of what makes a cluster and the operations would essentially be within this "black box". Would it make sense to add most of these notions to the BES discussion and the scheduler aspects would fit in RSS (where they could provide a basic RSS interface)? Putting these together with some minor additional specs may be the basic HPC profile. Am I missing something? Thanks! Ravi ________________________________ From: owner-ogsa-wg@ggf.org [mailto:owner-ogsa-wg@ggf.org] On Behalf Of Marvin Theimer Sent: Monday, March 13, 2006 2:42 PM To: Ian Foster; ogsa-wg@ggf.org Cc: Marvin Theimer Subject: RE: [ogsa-wg] Paper proposing "evolutionary vertical design efforts" Hi; Ian, you are correct that I view job submission to a cluster as being one of the simplest, and hence most basic, HPC use cases to start with. Or, to be slightly more general, I view job submission to a "black box" that can run jobs - be it a cluster or an SMP or an SGI NUMA machine or what-have-you - as being the simplest and hence most basic HPC use case to start with. The key distinction for me is that the internals of the "box" are for the most part not visible to the client, at least as far as submitting and running compute jobs is concerned. There may well be a separate interface for dealing with things like system management, but I want to explicitly separate those things out in order to allow for use of "boxes" that might be managed by proprietary means or by means obeying standards that a particular job submission client is unfamiliar with. I think the use case that Ravi Subramaniam posted to this mailing list back on 2/17 is a good one to start a discussion around. However, I'd like to present it from a different point-of-view than he did. The manner in which the use case is currently presented emphasizes all the capabilities and services needed to handle the fully general case of submitting a batch job to a computing utility/service. That's a great way of producing a taxonomy against which any given system or design can be compared to see what it has to offer. I would argue that the next step is to ask what's the simplest subset that represents a useful system/design and how should one categorize the various capabilities and services he has identified so as to arrive at meaningful components that can be selectively used to obtain progressively more capable systems. Another useful exercise to do is to examine existing job scheduling systems in order to understand what they provide. Since in the real world we will have to deal with the legacy of existing systems it will be important to understand how they relate to the use cases we explore. In the same vein, it will be important to take into account and understand other existing infrastructures that people use that are related to HPC use cases. I'm thinking of things like security infrastructures, directory services, and so forth. From the point-of-view of managing complexity and reducing total-cost-of-ownership, it will be important to understand the extent to which existing infrastructure and services can be reused rather than reinvented. To kick off a discussion around the topic of a minimalist HPC use case, I present a straw man description of such below and then present a first attempt at categorizing various areas of extension. The categorization of extension areas is not meant to be complete or even all that carefully thought-out as far as componentization boundaries are concerned; it is merely meant to be a first contribution to get the discussion going. A basic HPC use case: Compute cluster embedded within an organization. * This is your basic batch job scheduling scenario. Only a very basic state transition diagram is visible to the client, with the following states for a job: queued, running, finished. Additional states -- and associated state transition request operations and functionality -- are not supported. Examples of additional states and associated functionality include suspension of jobs and migration of jobs. * Only "standard" resources can be described, for example: number of cpus/nodes needed, memory requirements, disk requirements, etc. (think resources that are describable by JSDL). * Once a job has been submitted it can be cancelled, but its resource requests can't be modified. * A distributed file system is accessible from client desktop machines and client file servers, as well as compute nodes of the compute cluster. This implies that no data staging is required, that programs can be (for the most part) executed from existing file system locations, and that no program "provisioning" is required (since you can execute them from wherever they are already installed). Thus in this use case all data transfer and program installation operations are the responsibility of the user. * Users already have accounts within the existing security infrastructure (e.g. Kerberos). They would like to use these and not have to create/manage additional authentication/authorization credentials (at least at the level that is visible to them). * The job scheduling service resides at a well-known network name and it is aware of the compute cluster and its resources by "private" means (e.g. it runs on the head node of the cluster and employs private means to monitor and control the resources of the cluster). This implies that there is no need for any sort of directory services for finding the compute cluster or the resources it represents other than basic DNS. * Compute cluster system management is opaque to users and is the concern of the compute cluster's owners. This implies that system management is not part of the compute cluster's public job scheduling interface. This also implies that there is no need for a logging interface to the service. I assume that application-level logging can be done by means of libraries that write to client files; i.e. that there is no need for any sort of special system support for logging. * A simple polling-based interface is the simplest form of interface to something like a job scheduling service. However, a simple call-back notification interface is a very useful addition that potentially provides substantial performance benefits since it can enable the avoidance of lots of unnecessary network traffic. Only job state changes result in notification messages. * There are no notions of fault tolerance. Jobs that fail must be resubmitted by the client. Neither the cluster head node nor its compute nodes are fault tolerant. I do expect the client software to return an indication of failure-due-system-fault when appropriate. (Note that this may also occur when things like network partitions occur.) * One does need some notion of how to deal with orphaned resources and jobs. The notion of job lifetime and post-expiration garbage collection is a natural approach here. * The scheduling service provides a fixed set of scheduling policies, with only a few basic choices (or maybe even just one), such as FIFO or round-robin. There is no notion, in general, of SLAs (which are a form of scheduling policy). * Enough information must be returned to the client when a job finishes to enable basic accounting functionality. This means things like total wall-clock time the job ran and a summary of resources used. There is not a need for the interface to support any sort of grouping of accounting information. That is, jobs do not need to be associated with projects, groups, or other accounting entities and the job scheduling service is not responsible for tracking accounting information across such entities. As long as basic resource utilization information is returnable for each job, accounting can be done externally to the job scheduling service. I do assume that jobs can be uniquely identified by some means and can be uniquely associated with some principal entity existing in the overall system, such as a user name. * Just as there is no notion of requiring the job scheduling service to track any but the most basic job-level accounting information, there is no notion of the service enforcing quotas on jobs. * Although it is generally useful to separate the notions of resource reservation from resource usage (e.g. to enable interactive and debugging use of resources), it is not a necessity for the most basic of job scheduling services. * There is no notion of tying multiple jobs together, either to support things like dependency graphs or to support things like workflows. Such capabilities must be implemented by clients of the job scheduling service. Interesting extension areas: * Additional scheduling policies o Weighted fair-share, ... o Multiple queues o SLAs o ... * Extended resource descriptions o Additional resource types, such as GPUs o Additional types of compute resources, such as desktop computers o Condor-style class ads * Extended job descriptions (as returned to requesting clients and sys admins) * Additional classes of security credentials * Reservations separated from execution o Enabling interactive and debugging jobs o Support for multiple competing schedulers (incl. desktop cycle stealing and market-based approaches to scheduling compute resources) * Ability to modify jobs during their existence * Fault tolerance o Automatic rescheduling of jobs that failed due to system faults o Highly available resources: This is partly a policy statement by a scheduling service about its characteristics and partly the ability to rebind clients to migrated service endpoints * Extended state transition diagrams and associated functionalities o Job suspension o Job migration o ... * Accounting & quotas * Operating on arrays of jobs * Meta-schedulers, multiple schedulers, and ecologies and hierarchies of multiple schedulers o Meta-schedulers * Hierarchical job scheduling with a meta-scheduler as the only entry point; forwarding jobs to the meta-scheduler from other subsidiary schedulers o Condor-style matchmaking * Directory services o Using existing directory services o Abstract directory service interface(s) * Data transfer topics o Application data staging * Naming * Efficiency * Convenience * Cleanup o Program staging/provisioning * Description * Installation * Cleanup Marvin. ________________________________ From: Ian Foster [mailto:foster@mcs.anl.gov] Sent: Monday, February 20, 2006 9:20 AM To: Marvin Theimer; ogsa-wg@ggf.org Cc: Marvin Theimer; Savas Parastatidis; Tony Hey; Marty Humphrey; gcf@grids.ucs.indiana.edu Subject: Re: [ogsa-wg] Paper proposing "evolutionary vertical design efforts" Dear All: The most important thing to understand at this point (IMHO) is the scope of this "HPC use case," as this will determine just how minimal we can be. I get the impression that the principal goal may be "job submission to a cluster." Is that correct? How do we start to circumscribe the scope more explicitly? Ian. At 05:45 AM 2/16/2006 -0800, Marvin Theimer wrote: Enclosed is a paper that advocates an additional set of activities that the authors believe that the OGSA working groups should engage in. Broadly speaking, the OGSA and related working groups are already doing a bunch of important things: * There is broad exploration of the big picture, including enumeration of use cases, taxonomy of areas, identification of research issues, etc. * There is work going on in each of the horizontal areas that have been identified, such as EMS, data services, etc. * There is working going around individual specifications, such as BES, JSDL, etc. Given that individual specifications are beginning to come to fruition, the authors believe it is time to also start defining vertical profilesthat precisely describe how groups of individual specifications should be employed to implement specific use cases in an interoperable manner. The authors also believe that the process of defining these profiles offers an opportunity to close the design loopby relating the various on-going protocol and standards efforts back to the use cases in a very concrete manner. This provides an end-to-end setting in which to identify holes and issues that might require additional protocols and/or (incremental) changes to existing protocols. The paper introduces both the general notion of doing focused vertical design effortsand then focuses on a specific vertical design effort, namely a minimal HPC design. The paper derives a specific HPC design in a first principlesmanner since the authors believe that this increases the chances of identifying issues. As a consequence, existing specifications and the activities of existing working groups are not mentioned and this paper is not an attempt to actually define a specifications profile. Also, the absence of references to existing work is not meant to imply that such work is in any way irrelevant or inappropriate. The paper should be viewed as a first abstract attempt to propose a new kind of activity within OGSA. The expectation is that future open discussions and publications will explore the concrete details of such a proposal. This paper was recently sent to a few key individuals in order to get feedback from them before submitting it to the wider GGF community. Unfortunately that process took longer than intended and some members of the community may have already seen a copy of the paper without knowing the context within it was written. This email should hopefully dispel any misconceptions that may have occurred. For those people who will be around on for the F2F meetings on Friday, Marvin Theimer will be giving a talk on the contents of this paper at a time and place to be announced. Marvin Theimer, Savas Parastatidis, Tony Hey, Marty Humphrey, Geoffrey Fox _______________________________________________________________ Ian Foster www.mcs.anl.gov/~foster Math & Computer Science Div. Dept of Computer Science Argonne National Laboratory The University of Chicago Argonne, IL 60439, U.S.A. Chicago, IL 60637, U.S.A. Tel: 630 252 4619 Fax: 630 252 1997 Globus Alliance, www.globus.org <http://www.globus.org/>
participants (2)
-
Marvin Theimer -
Subramaniam, Ravi