Introduction
After a pair of Citrix NetScaler SDXs to buy, they rack mounting and it is powered up, which is the next step?
, the next steps would be to always access the management instance that create "service Virtual Machine" (SVM) and start virtual NetScaler instances on the physical SDX platforms.
for the best elasticity, it is wise to have two physical SDX platforms within a data center and a few virtual instances in a high availability arrangement pair (they are so active / backup for the workload) actively in the two separate physical systems. The instance cluster option is now also available should an active / active configuration may be required.
This is affecting a non-service design, which can be bent to provide service as needed. This can then be made available to high above sites with something like Global Server Load Balancing.
Figure 1 A NetScaler SDX, with a single virtual instance
Regarding a quick start process there are documents online, describing how new create instance, when it is possible to have the review of the notes read here:
http://docs.citrix.com/en-us/sdx/11/configuring-managing-netscaler-instance.html
The purpose of this document is to provide guidance as to a SDX system to dismember the resources available, and what to create the various options for a new instance (s).
, the workload itself for the new instance is required, should also be considered when dimensioning the estimate because different workloads have different requirements from the platform.One the main advantages of the SDX is the to revise option allocation as needed, so to some extent it would be possible to make a rough estimate of first and to revise this up or down according to need
. Note: This document relates to the entities that exist on the SDX be specific rather "virtual instances" as a "NetScaler VPX". Citrix NetScaler VPX also has a product to sell, which is developing a product for a number of commercial hypervisor. As a result, with respect to the available power, the virtual instance on the SDX very different from that on other platforms. A SDX virtual instance is a much higher burden capable
Assumptions
It is assumed that
- The reader understands. The allocation of IT resources in a hypervisor-based system.
- , the reader has some familiarity with the NetScaler Application Delivery Controller.
- , the reader has some familiarity with Citrix software products.
NetScaler SDX.
| SDX system | instances in come: models There are currently six SDX systems available for sale, they are in following summarized device | maximum number of instances | throughput | Platform limit |
| 8015 system | 2 | 5 | 15Gbps | 15Gbps |
| 11515 series | 5 | 20 | 15Gbps | 42Gbps |
| 14020 series | 5 | 25 | 20Gbps | [1945024100Gbps]|
| 17550 series | 5 | 40 | 20Gbps | 50Gbps |
| 22040 series | 20 | 80 | 40Gbps | 0Gbps |
| 24100 series | 40 | 80 | 100Gbps | 150Gbps |
Table 1 SDX models
the border platform is the ability to have the platform for throughput flexibility, so how much more can the return transfer to the appropriate platform license. As can be seen from the table, this system have very flexible and thus an excellent starting point to make ADC density for the drive and so costs for things such as maintenance saving.
When a SDX buying there are a few ways to add instances, these packs can be added with 3 or 5, as soon as the packages, it is only a matter of resources division added to define what the number of is associated with instances that are available.
How resources assigned Virtual instances
inside the SVM management console contains a series of tabs that can be used to check what the status of SDX appliance system currently, the election of the Dashboard tab, SDX resources are considered accepted.
The following screenshot shows an old 20500 platform with 3 x 5-instance-packs and the relative throughput to new available (this is a 50Gbps appliance) and that the flow rate, which remains (only 33Gbps throughput) for mapping instances. It should be noted that when a virtual instance most of these funds are created permanently assigned, generally is no overcrowding of resources in SDX. However there are some exceptions to this rule.
Figure 2 20500 SDX Resources
In the above example, the SDX has eight virtual NetScaler instances that have created. Some are active and some are below. There are a number of ways when an instance to create, CPU, memory, throughput and SSL key task. So the SDX is a series of virtual instances hosting 17GB throughput totaling 7 SSL cores for SSL processing assigned along with 24 GB of system memory in this case.
These values were assigned using a wizard that is located under the "Configuration" tab. This tab shows the various controls to configure the structure and SDX appliance. The wizard can be used to create new instances and reconfigure existing instances.
For example, after the base platform Stock Exchange more platform throughput with a simple software key, give them the resources might be necessary to revise available a running virtual instance. Add The revision might or remove throughput to an instance, which must manage whatever the administrator available resources and workloads.
The following screenshots of Provisioning Wizard show assign these values, these provisioning wizard for each SDX system. The only difference between the models, the resources and the flow rate for the assignment.
An indication of the Docs page points out that most of the changes do not require that the instance is restarted. However, if the following parameters are changed :. Number of SSL chips, interfaces, memory and feature license, the NetScaler instance implicitly stops and starts these parameters to bring into force
| step | Description | background |
| 1. | [assign instance name, management address data, feature set, Administration profile and a description. | is contain basic information, for example the option to provide a Standard, Enterprise or Platinum feature license, if that is all that is required. |
| 2. | of available resources to assign, so in this case, memory, SSL cores, flow mapping mode (fixed or burstable), priority minimum throughput (MBps) maximum burst rate (mbps), burst, PPS and CPU affinity and core task. | , the ability to assign to the instance for a particular request. |
| 3. | creating an instance specific administrative account. | A special account as management. |
| 4. | illustration of the network connection, for example, | This allows the example of the network connection of the SDX to have basic platform. |
| 5. | The final step confirms the information and agrees that changes when an invalid setup it was requested to be marked by the wizard at this point. | process is finished, errors must be corrected before being committed with target. |
Table 3 Instance Creation
Available resources and the allocation of
In step 2 of the previous section, we had this screenshot:
Figure 3 instance sizing
This is the important part of the virtual instance wizard in terms of resource allocation, as that is where its relative performance is defined.
Here is a table showing the resources available per SDX system.
| SDX system parameters | 8015 System | 11515-11542 systems | 14020-14100 systems | 17550-21550 systems | 22040-220 systems | 24100.24150 systems |
| CPU type | 1 x four core | [19450422] x Six core2 x Six core | 2 x Six core | 2 x eight-core | 2 x eight-core | |
| The available memory | 32GB | 48GB | 64Gb | 96GB | 256GB | 256GB |
| SSL cores | 4 | 16 | 16 | 36 | 128 | 16 |
| throughput | 15Gbps | 15-42Gbps | 20-100Gbps | [194504220-50Gbps]40-0Gbps | 100-150Gbps | |
| CPU cores including management | 4 | 12 | 12 | 12 | 16 | 16 |
| CPU cores for allocation [availableare | 3 | 10 | 10 | 10 | 14 | 14 |
Table 4 A summary of the SDX platform functions
the SDX system has only two limits, throughput and instances; everything else is purely a function of these values. The following sections will detail the specifics about each resource type and how they are related.
It is worth considering that the SDX is a resource pool, in the allocation of resources decisions, the available resources can change the left are assigned to new instances. The sizing metrics for Saying SDX11515 quote this 20 instance system with its full allocation of instance-pack.
If the allocation of resources, the selected options can mean below that this is something less than 20 instances for a particular application. Here are the details of the various options available:
memory allocation
| SDX system parameters | 8015 system | 11515-11542 systems | 14020-14100 systems | 17550-21550 systems | 22040-220 systems | 24100.24150 systems |
| The available memory | 32GB | 48GB | 64Gb | 96GB | 256GB | 256GB |
Table 5 the available memory for SDX platforms
type of allocation? is difficult assigned this resource, it is not possible to commit this resource on the SDX to.
On each SDX platform the size of the memory for the entry version is the same as it is for the top version in a particular series. So that. 11,515 ships with 48GB of memory and this is available, purchased regardless of version (11515 or 11542) The SVM needs a small amount of memory to work, so that they can to 4Gb less than the memory installed in the system for mapping instances.
There are some limitations with respect to SSL cores. Memory and SSL cores are connected to each other so that when each SSL core it will be necessary to assign, 1 GB to be assigned per core memory.
SSL cores
| system parameters | 8015 System | 11515-11542 systems | 14020-14100 systems | 17550-21550 systems | 22040-220 systems | 24100.24150 systems |
| SSL cores | 4 | 16 | 16 | 16 | 128 | 16 |
Table 6 SSL Scores for SDX platforms
type of allocation? is difficult assigned this resource, it is not possible to commit this resource on the SDX to. As already mentioned, the SSL core and the memory allocation in connection.
The actual number of SSL cores depends on the allocation on the platform license that has been purchased, although SSL performance is based on let throughput, more throughput faster processing of SSL traffic.
There is an option to create a virtual instance, the assigned no dedicated SSL cores, software-based SSL must be used in this case. Software-based SSL is obtained processed by the CPU. Consequently, it may have less overall performance compared to an instance with special SSL key (s) again. Depending on the traffic profile
Throughput
| system parameters | 8015 System | 11515-11542 systems | 14020-14100 systems | 17550-21550 systems | 22040-220 systems | 24100.24150 systems |
| throughput | 15Gbps | 15 - 42Gbps | 20-100Gbps | 20-50Gbps | 40-0Gbps | 100-150Gbps |
Table 7 throughput for SDX platforms
type of assignment? This can be a resource to be assigned or set hard breakable; The flow rate is measured as the sum of the received traffic in the NetScaler. exceeds a certain level, the unit will fall trigger traffic on the fixed setting. When using the tearable option options available commitment rate over the target throughput value is to allow, via required. The burst priority then sets which instance prevail when two instances both need a set amount blurted.
In order clear the "burst" is the value to the minimum flow rate, the maximum to offer in addition to the instance flex to up. Billing metrics are used to track how often this happens, so allows the use of track. assign
Figure 4 burst on an instance tracking
Regarding the required throughput per instance, will determine the entire platform license, which is available before instances created be, so this could be 15GB or 150Gb, for example, with this pay-as-you-grow may change with the specific SDX, was obtained.
CPU core allocation
| system parameters | 8015 System | 11515-11542 systems | 14020-14100 systems | 17550-21550 systems | 22040-220 systems | 24100.24150 systems |
| CPU cores including management | 4 | 12 | 12 | 12 | 16 | 16 |
type of allocation? This really depends on the options selected in the wizard, if the resources are assigned to this hard, the number of cores is reduced, which can be shared between the remaining instances in the SDX system. All cores are designed for a specific platform for all licenses for this platform.
With respect to the number of cores for common instances. There is no limit a single core for shared instances assign. Instead, the device all available cores to use instances attempts (which were assigned for dedicated instances not) for shared instances and distributed over these cores.
So if you have 10 physical cores on the device and 10 divided create instances, each shared instance gets its own physical core. But if the number of common instances start, the number of available cores to pass, then the same physical core is shared among multiple instances. So, in the above example 11. Example divides a physical core with the first example, the twelfth to the second, and so on.
In short, the device follows a best fit algorithm for the common cases, based on the number of available cores. This gives the user the ability to mix and match dedicated and shared instances. For example, the user may 5 dedicated instances with a core provide each and 10 shared instances over the remaining five cores are assigned.
The assigned CPU cores is directly related to the number of packet engines that are assigned (maximum) running on the virtual instance on the SDX appliance.
Figure 5 SDX core task
Here is a screenshot of the option to set processor affinity (see above). This screenshot of a 11500-Appliance, the actual number of cores was taken, which can be assigned will vary on the processors in the device. Larger plants (200 systems, for example) is at 7 dedicated cores allow up to assigned.
It should be noted that when cores are dedicated to virtual instances, the number of instances that can be reduced can be created. This is because when the nuclei have a 1: 1 relationship with a virtual instances they are no longer available to be shared between instances
Note1 - core allocation can not CPU socket on platforms extend with. multiple physical CPUs. Core allocation is thus dependent on prior core allocations and available cores on a per physical CPU basis
Note2 -. The cores mentioned in this document are the physical cores that are available in the system. Hyper-Threading will mean that the SDX24040 have 32 cores appears for example. This screen shot below shows this.
Summary
This document describes the various options for the allocation of resources for one NetScaler SDX has virtual instance defined, this is purely to provide a good first make estimates for the instance. , Re-running the Provisioning Wizard allows dynamic changes are applied without compromising Live service
More information is available online at: http://docs.citrix.com/en-us/sdx/11/ SDX-introduction.html
0 Komentar