Gigaspaces.com - Application Server
Home
  • Send to a friend
  • Print

Webinar's Q&A - Getting More out of Virtualization Faster with XAP

  • What is the role of GS-Agent in loading additional virtual machines/GSCs?
  • A: The GS-Agent is a process which runs on every virtual machine and is responsible for controlling the life-cycle of every GigaSpaces component on this virtual machine. The GS-Agent is not, however, in charge of starting or stopping virtual machines. This procedure is done via the VMWare APIs; currently the vCenter API - and once it’s matured, the vSphere cloud APIs.
  • Q: What is the difference between GS Agent and GSM?

  • A: The GS-Agent is in charge of the life cycle of all GigaSpaces components on a single machine/VM.

    The GSM is responsible for the application life-cycle, deployment and undeployment of processing units and web applications. More information.

  • Q: We anticipate that the overhead of VMware is expensive and uses too much CPU for CPU intensive low latency operations.
  • A: Although optimized much in last version, VMWare hypervisor technology still holds a concept of multiple virtual machines who share the same physical resources. GigaSpaces layer allows the user to control the deployment environment: the number of processes of on each virtual machine, the number of threads that are used and also the amount of memory that is being allocated. In addition, some high performance applications will be deployed directly on the physical machines, without any change to the application architecture.
  • Q: When the Database is down, how does an application access the non-loaded data (e.g. lazy loading)?

  • A: When using the topology of all-in-cache and the database is updated asynchronously in the background, there is no need to access the database during the regular life-cycle of the application. The database is accessed when the system first starts and is being updated asynchronously to ensure consistency with the information stored in the spaces memory.

    Choosing a different topology, where critical information is stored in the database and accessed on demand, requires a highly-available database in order to enable an on demand high-availability of the system.

    This is, of course, a less recommended solution. A better procedure is to use applications who can store all the information within the space control (memory and/or disk).

  • Q: Can you auto scale a partitioned data grid in 7.0.1?
  • A: The auto scaling capabilities which were demonstrated as part of the webinar will be part of the new offering. These are available today as part of R7.0.1 by using the Admin API.
  • Q: What is the new support on indexes and joins on the space?
  • A: In Release 7.0/7.0.1 the product index handling and storage was improved dramatically. In addition a new capability for dynamic indexing is being added in to the current development stream.
  • Q: Is it possible to setup a clustered filesystem and scale it at the same time as the rest of the cloud?
  • A: Yes it is. There are some projects that have built a distributed file-system on top of GigaSpaces XAP. One of them is Compass the open source java search project.