License sharing is used by software vendors to allow multiple concurrent uses of a single license.
Layman's article about the rationale and purpose of heartbeats.
Adding support for license queuing promotes more efficient use of a user’s inventory of licenses
The RLM license server has the ability to manage a queue of requests for a license when all licenses are in use. If queued for a license, the application is notified by the license server when a license has become available. Queuing occurs in the course of a normal checkout if the application’s environment specifies a value for the variable RLM_QUEUE.
Applications must be made aware of Queuing
The licensed application needs to support queuing, in that it has to react appropriately to the RLM_EL_INQUEUE error status after a checkout attempt, instead of treating it as a hard error. The application should keep checking the license status periodically as long as the user wishes to remain queued for the license. Not all RLM-licensed applications do this, but most do.
Queuing to Multiple Servers is Transparent
The licensed application side of RLM manages queuing in a multiple license server environment transparently. If all licenses are in use on all of the license servers that the licensed application is aware of, RLM will queue for a license on all of the license servers. As soon as a license becomes available on one of the servers, it is granted to the queued application, the application dequeues on the other servers. This is handled transparently to the application, in the RLM library. [ Don’t you wish supermarket check-out lines worked this way? You’d be waiting in all of the lines at once. ]
All Aboard the Express
RLM has the concept of “express” license pools for the purposes of queuing. An express license pool is one where queuing is not on a strictly first-in first-out basis. This is best illustrated by example. Consider a pool of 10 licenses: 8 are currently in use, and 2 are available. Application A requests a checkout of 3 licenses, and specifies that it wants to queue if the request can’t be granted immediately. Its request is queued because only 2 licenses are available. Application B then requests 2 licenses. If the pool is an express pool, Application B’s request is granted immediately. In a non-express pool, Application B’s request would be queued behind Application A’s request. By default, all pools are express pools. This can be changed by the end user on a pool by pool basis, or globally, via the EXPRESS directive in the ISV options file.
Software developers will freely admit that they are not licensing experts. When they asked to evaluate, select and implement a software licensing solution they want one that is robust, yet quick to implement and easy to maintain.
RLM end-users remind me of why software license management is so important to them.
by Cody Crider, Reprise Software, Inc.
We have written many articles about the benefits of software licensing to Independent Software Vendors (ISVs) and how to implement various features within RLM. Of course this is the target audience for Reprise Software’s product, right? That might be where we make our living, but there is a whole other world of software licensing that ISVs should examine more closely, from their customers’ perspective.
I recently had the opportunity to make a presentation to a number of end-users of enterprise-class software products that utilize software licensing including the Reprise License Manager (RLM). These end-users represent some of the largest companies in the world. It was very enlightening to hear about software licensing from the perspective of the folks involved in the day-to-day management of hundreds of applications, and thousands of licenses. The special value of software licensing to this end-user community is one that needs to be continually addressed.
There were the discussions about what end-users commonly want, such as the following:
- Better control over when/where/which users can access software assets
- Unfettered access to usage data and tools to track and allocate costs
- Improved license administration tools
- More predictable and consistent licensing models across vendors
- Transparent policy spelled out in the license key
- Auto-maximized license utilization
- Better control over license queuing
- Better performance measurement tools to plan for and deploy new licensed applications
- Better diagnostic tools
- Simpler Internet-based activation and license installation
We had an opportunity to discuss how RLM address many of the needs outlined above and that the majority of the innovations in RLM are in support of better value for end users while simplifying implementation and deployment for ISVs.
The discussion then shifted a bit when the question was asked, “Does RLM do a good job at ensuring software compliance?” Sure, we talk about a benefit of RLM is to ensure that ‘honest users stay honest’, but this is from the perspective of the ISV who wants to maximize revenues by eliminating unlicensed usage. A light came on to remind me that this is also a critical issue for our end-user community. From their perspective, it is one of the benefits they value most from software licensing. Let’s face it. There are a lot of things that end-users don’t like about software licensing, but it is one of the ‘inconveniences’ that they can tolerate for all the extra benefits derived from it.
The software end-user has a management mandate to stay compliant with their software licensing agreements and to have the information they need to prove this during audits. Software compliance can be less costly for these customers if the software they license includes the proper built in license compliance tools, like RLM.
RLM provides the flexible licensing models and features that can match what is being sold to software end-users. ISVs who support license models matching those described in their software licensing agreement can go a long way to help ensure compliance.
RLM provides the administration tools to monitor and support license compliance reporting. The tools are built into RLM so that the end-user can easily control how the software is utilized in their organization and determine who has the rights to use it. RLM provides detailed license usage report logs that can be used to track and monitor usage across the organization. The RLM Reportlog uses an open and authenticated format that can easily be read for reporting purposes. Reprise works closely with 3rd party vendors such as License Tracker, OpeniT, and RunTime Design Automation that provide tools to help maximize the usage and monitoring of software assets.
All of these features and efforts can help assist the end-user community to ensure that they are in compliance with their license agreements.
Advanced End User RLM Administration Tutorial Series
Reprise Software continues its series of regular on-line tutorials designed to help users maximize the value of the Reprise License Manager (RLM). These sessions are designed for system administrators and ISV tech support people.
This third session will be conducted by Reprise Software’s lead software developer, Bob Mearns, who has over 15 years experience developing and supporting license manager software.
The 40-minute session is scheduled for Tuesday October 12th at 8AM PDT (GMT-08). This session will focus on RLM license server administration and optimization in a multi-server environment. It will specifically address:
License server administration of multi-server sites
- Connecting to multiple license servers
- Connection and Idle Timeouts
Server to Server License Transfers
- License Queuing
- RLM’s environment variable settings
For more information about the tutorial series, please contact Mr. Cody Crider at [email protected]
Technical article on timeouts in the Reprise License Manager
There are two types of timeout events in RLM: connection timeouts, and idle timeouts. Let’s dig into each of them a little bit deeper.
Connection timeouts occur when a client application takes too long to connect to one or more RLM license servers. Users can change the default connection timeout value from 10 seconds to as low as 5 seconds by setting an environment variable. If set to a negative value, the connect timeout will be its absolute value, and if any particular server connection times out, no further attempts will be made to that server. If set to a positive value, a connection will be attempted to the server even if it timed out on the last attempt. The latter example is the default behavior in RLM.
Skipping over License Servers
RLM also keeps track of which license servers in its license path list have an ISV’s server running on them. If either there is no ISV line for that ISV in a local license file, or an attempt is made to check out a license from a server and the RLM server returns a status indicating that this ISV is not present, then the server is flagged as not having that ISV server.
In a multi-server environment, it doesn’t make sense for an application to repeatedly attempt checkouts from servers that are known not to be serving its licenses. So, application developers may choose to “skip” these license servers. Note that by default, RLM will attempt all operations on all servers.
Idle Timeouts and Heartbeats
Most users want the license server to reclaim licenses from applications that become idle for too long. This ensures maximum utilization of floating (shared) license pools.
The RLM license server expects to hear from client applications periodically as a way for applications to tell the server that they are still active. Normally, this is done by sending “heartbeats” from clients to the server either automatically or under the direct control of the application itself.
Some software vendors may decide to inform the server that an interactive application is idle while waiting for user input. This allows the application’s license to be reclaimed by the server when the application waits too long for user input. When the user eventually “wakes up” the formerly idle application, it will attempt to reclaim any lost licenses from the server.
When licenses are timed out by the server, they are essentially checked back into the pool for use by others. This event can be recorded in a server report log, if desired.
User Control of Idle Timeouts
An end-user can specify a TIMEOUT value for any idle license. If the license remains idle for this amount of time, the license server performs an automatic check-in of the license and informs the application (if it is still running).
Software vendors can set a MIN_TIMEOUT keyword in a license that specifies the lowest value that an end-user can set for the timeout value of a license. If not specified in the license, the RLM default minimum of 3600 seconds (1 hour) is used.
Sometimes, it takes a customer to prod you into a new way of doing business to help you understand the full functionality that you've had all along.
Consider your options
The usual licensing strategy for server software is to lock it to a physical host. This provides some degree of protection from installing and running on multiple hosts, but it does not necessarily represent the optimal approach in terms of maximizing your revenue or customer satisfaction.
First, there’s the notion of scaling. Do you really want to charge the same price for every site? You need some way to charge more when your software is used more. Perhaps you can sell multiple node locked servers to larger sites, or even charge relative to the number of cores on the machines, the rationale being that there is the potential to get more use from machines with more horsepower.
Float your server licenses
If you want your customers to have the flexibility of installing and using your server on many machines without having to license them separately, perhaps you should use a floating license manager that ensures that no more than a predetermined number of concurrent server copies can run at the same time. With floating licensing, you issue a license key with a count, and lock only the license server, not each user node.
Another approach could be licensing by some other scaled metric: how many records can be maintained in the database, a count of named users, etc. Each of these metrics can be encoded into a license key, allowing the server software to know when it starts what level of service to enforce based on the license key scaling parameters. A single version of your software could then enforce the appropriate size limits on a site by site basis.
Using a license manager, such as RLM, could also give you even more control of how your software behaves within a virtual machine. You could even license by platform or by the site’s timezone.
In any case, using a license manager gives you the freedom to change with your license policies.
In an attempt to clear up some confusion and provide a little education on the not-quite-fascinating subject of license servers, we'll go over various licensing scenarios and what is involved, server-wise.