Frequently Asked Questions
Q:
I get an error message saying the license is not valid or missing? What do I do?
Answer
The Aleri stream processor, the run time engine of the Aleri Streaming Platform, requires a valid license key to run. The license key is contained in a file called license.key that must be present in the .../AleriStreamingPlatform/etc directory. You can run the Aleri Studio without a license key, but if you try to start the stream processor, you will get an error message if the license key is not present or has expired. If you downloaded the Aleri Streaming Platform from the Aleri website, you should have also downloaded a license key file. It must be downloaded separately from the product install package. If you did not download the license key, you can go back to the download section of aleri.com, logon, and download it. Once you have downloaded the key, install it per the instructions. Note that only evaluation licenses can be downloaded from aleri.com. Customers who have purchased the Aleri Streaming Platform will be given a production license. If you are an Aleri customer and have not received your production license or are having a problem with your license.key file, contact Aleri Customer Support at support@aleri.com or your Aleri representative.
Q:
How do I install a license key?
Answer
The license key is contained in a file called license.key. Simply place this file in the .../AleriStreamingPlatform/etc directory of the machine that the Aleri server will be running on. If you installed the Aleri Streaming Platform on a Microsoft Windows PC and used the default install location, this will be C:\Program Files\Aleri\AleriStreamingPlatform\etc. If you are installing a license extension, simply replace the old license.key file with the new one. The next time the sever is started it will read the new license key file.
Q:
How do I install and run the Aleri Dashboard?
Answer
A: The Aleri Dashboard is an optional add-on to the Aleri Streaming Platform. It is not included in the evaluation package that can be downloaded from aleri.com. We are able to provide evaluation licenses for the Aleri Dashboard; if you wish to include the Aleri Dashboard in your evaluation of the Aleri product, contact Aleri to make arrangements.
Q:
When running the Studio on Fedora Core 5, I can't use the Ctrl key to select two or more options from a list (for example, a set of indices). How do I fix this?
Answer
To enable this feature:
1. In the Fedora Core GUI, select System > Preferences > Mouse.
2. In the panel that appears, deselect HIGHLIGHT MOUSE.Q:
What Operating Systems is the Aleri Streaming Platform supported on?
Answer
Currently the Aleri Platform is supported on Solaris 10, Linux (RedHat and Suse), and Microsoft Windows. Other operating systems/versions may be added in the future, so feel free to discuss your specific requirements with your Aleri representative.
Q:
Can I run the Aleri Studio on my desktop and connect to an instance of the Aleri event processor running on a remote server?
Answer
Absolutely. The Aleri Studio includes features for remote execution. For full use of all features in the studio, this configuration does require that the desktop and server both have access to a shared disk where the data model can be edited and accessed. See the Installation Guide for instructions on how to configure the studio for remote execution.
Q:
How much RAM does the Aleri Platform require?
Answer
The amount of RAM is determined by the amount of data that will be retained in the data model. Individual streams in the data model are configured as stateless or state-full. The amount of data held in a stream that retains state is determined by the size of the retention window which can be based on time or number of records. Thus the amount of RAM required is a function of how much data will be held within the event processor when all retention windows are at their maximum.
Q:
What is the Aleri Dashboard used for?
Answer
The Aleri Dashboard is used to build visually rich real-time displays of output from the Aleri Streaming Platform. The dashboard builder makes it easy to build custom dashboards in a matter of hours, selecting from a rich palette of pre-built display objects. The Aleri Dashboard allows the Aleri Streaming Platform to be used for Operational Intelligence and Business Activity Monitoring (BAM) applications.
Q:
Can I run multiple data models on the same server?
Answer
A single server (machine) can run multiple instances of the Aleri event processor. Each instance of the Aleri event processor runs a single data model, but a single data model can have any number of streams. Thus, if there is a need to run multiple data models on a single instance of the event processor, it’s relatively simple to combine multiple data models into a single large model – even if the individual models are unrelated.
Q:
What's an Aleri data model?
Answer
The Aleri data model contains all of the event processing logic that will be applied to input streams to produce results. You could think of this as the application that runs on the Aleri CEP engine. We don’t call it an application, since a complete application will involve not just the data model but the input data sources and the connections to downstream applications or dashboards that will use the results. An Aleri data model is created using one of the Aleri authoring environments: the Aleri Studio, Aleri SQL, or in XML following the Aleri schema.
Q:
When should I consider using a FlexStream in my data model?
Answer
FlexStreams give you the ability to apply procedural logic to an incoming event. This overcomes the limitations of a purely declarative language such as SQL. For many frequently used operations such as filters, aggregation, or computations on a single record, standard relational operators can be easily applied. There are cases, however, where you may need a greater level of control or may need to loop or iterate. This is when you will want to include a FlexStream in your model.
Q:
How many input streams can a single instance of the Aleri Platform support?
Answer
There is no absolute limit. The practical limit will be dictated by a variety of factors including: message rates, the amount of data to be retained within the Aleri server, number and speed of CPUs/cores, amount of RAM, etc
Q:
The product literature says that Aleri supports both High Availability and Clustered configurations. What's the difference?
Answer
In a High Availability configuration, two identical Aleri CEP engines run in parallel, one as a primary, the other as secondary. Both are live; all data is absorbed by both servers. Should the primary fail, the secondary will automatically take the primary role and all subscriptions are moved from the primary to the secondary. In a clustered configuration the data model is split over n servers/blades, these n servers/blades are designated as hot, and one or more additional servers/blades can be designated as spares. Should one of the hot servers or blades fail, one of the spares is automatically selected and brought online. The streams that were running on the failed server are automatically restarted on the new server. This mode requires a clustered file system on a RAID or SAN, because our disk based persistence is required to recover the data from the failed server/blade onto the newly activated server/blade. The recovery time in this mode is measured in seconds and this recovery time is dependent upon the data model and the amount of data that needs recovery.
Q:
What is a Continuous Query?
Answer
We often describe some of the operations within an Aleri data model as continuous queries. A database query in a traditional database environment applies a query to a set of data in the database to produce a result set. A continuous query uses the same type of query logic but inverts the process: the continuous query is defined ahead of time and then as data arrives it passes through the continuous query to immediately and continuously update the result set. So the input to a traditional query is one or more tables and the output is a view (or materialized view). The input to a continuous query is one or more streams and the output is a derived stream.
Q:
How many messages per second can the Aleri Platform process?
Answer
Quantifying throughput (messages per second) is difficult because there are so many variables involved. Aleri measures throughput as the number of incoming messages that can be processed in real-time without the Aleri server getting overloaded (which would result in queuing and delays). The main things that affect maximum throughput are: hardware (processor speed, number of processors/cores, amount of RAM, etc) and the complexity of the data model (the event processing logic being applied to the incoming data). Maximum throughput for a single server may range from as little as 20,000 messages per second to over a million messages per second depending on the hardware configuration and data model.
Q:
How much latency is introduced by the Aleri Platform?
Answer
Latency is largely a factor of the data model—i.e. the complexity of the event processing logic being applied and how much the computations can be parallelized. Hardware will of course also play a role in determining latency. Latency within the Aleri platform is typically sub-millisecond; latency on individual streams can be as low a few microseconds.
Q:
How does the Aleri Platform connect to external systems?
Answer
Aleri provides a range of adapters that are available for connecting the Aleri CEP engine to data sources and consumers. The base product includes an adapter for both reading data from files and writing data to files - in both cases csv and xml parsing is supported. Adapters are also available from Aleri for a range of data sources and consumers including: JMS, TIBCO, MQ, Reuters (Marketfeed and OMM), JDBC, ODBC and others. Our list is constantly growing; contact Aleri regarding the availability of specific adapters that you may need. In addition, the base product includes the Aleri pub/sub API, available for C++, Java, and .NET. This API can be used to build custom adapters or can be integrated directly with an application. Note: the product package that can be downloaded from the website only includes the adapter for input/output to/from files and the API; it does not include any other adapters. Specific adapters can be provided for evaluation, however, just contact Aleri to make arrangements.
Q:
How long does failover take in a High Availability configuration?
Answer
Failover is pretty much instantaneous. As soon as the secondary detects that its connection to the primary went down or it doesn't get a heartbeat message, it takes over as primary, hence sub-second. Sources publishing data into the platform that either run synchronously or do periodic commits will need to detect that the failover occurred (the API provides notification) and resubmit everything since the last commit.
Q:
Aleri offers XML as an event processing language. Is this standard XML and if so, how does it work?
Answer
Yes, it is 100% standard XML. Aleri has defined a schema that defines the elements and attributes that can be used within the data model. The XML document then defines each stream, the inputs to the stream (thereby defining the dataflow), the data structure of the stream (the row definition), and the operations that will transform the input stream(s) into an output stream.
