Managed Load Balancer

Menu Overview

Features/Benefits

The Managed Load Balancer menu provides your own L4 load balancer function. You can use the load balancing function that distributes external communications to multiple servers. (The equipment that provides the functions of this menu is called a Managed Load Balancer Instance.)

  • Managed load balancer

    • The service provider will operate, monitor, and recover from the failure in this menu.

    • You can concentrate on your business because you use the assets immediately without owning them, initial investment and minimum usage period.

  • Automatic provisioning

    • Managed Load Balancer creation, configuration, and deletion are automated, so you can flexibly use resources as much as you want, when you want.


Use Case

  • You can use the load balancing function to distribute communication from outside or inside to multiple servers.

  • The target group specification function enables load balancing of multiple systems with one device.

  • By operating the portal / API of the load balancer, you can easily manage and control the settings and status.

ManagedLB

Available Functions

List of Functions

The functions available in this menu are shown in the table below.

No.

Functions

Overview

1

Load balancing function

It provides load balancing functions such as health monitor, listener, and target group settings.

2

Network function

It provides a connection / disconnection function to a logical network and a static route function.

3

Instance control function

It provides the ability to create, edit, and delete Managed Load Balancer Instance.

4

Management function

It provides a monitoring function and GUI / API as functions for managing and operating the Managed Load Balancer.


Description of Functions

1.Load balancing function

As a load balancing function, it provides a setting function for health monitor, listener, target group, and policy (L4 load balancing algorithm, persistence methods).

Item

Contents

Health monitor

・It periodically monitors the failure of the load balancing target (real servers). When a failure occurs, it automatically disconnects them from the load balancing target.
・You can use the health check function at the L3 (ICMP) and L4 (TCP) levels.
・You can set the interval (seconds), the number of retries, and the value of timeout as the health check threshold.

listener

・This function accepts requests to connect to the Managed Load Balancer.
・You can specify the IP address, port number, and protocol.
・The protocols that can be used are [TCP, UDP].

Target group setting

・You can group the servers (members) to which the load is distributed.
・You can define the target group name and specify the member server by IP address and port.
・You can specify the load balancing weight in the target group for each member server.

Policy settings

・You can configure algorithms and persistence for L4 load balancing by specifying the Health Monitor / Listener / Target Group to.
・You can choose the load balancing algorithm from the following methods.
–round-robin: Sorts the requests to the server in order.
–weighted-round-robin: Sort requests based on the weight assigned to the server.
–least-connection: Distribute the request to the server with the fewest connections.
–weighted-least-connection: Sorts requests based on the combination of weight and number of connections assigned to the server.
–source-ip-port-hash: Distributes the request to the server based on the hash value of the source IP header and port number from the client.
・You can select [Condition source IP address method (source-ip) / No setting (None)] as the persistence (session maintenance method).
–none: Not set the persistence (session maintenance method).
–source-ip: Requests from the same source IP address are distributed to the same server.

Note

  • L7 load balancing and SSL offload functions are not provided.

  • There is an upper limit to the number of settings for the load balancing function, so please refer to the upper limit list in “Application Consideration”.

  • Only one health monitor / listener / target group can be specified in one policy.

  • It is not possible to listen to both of TCP and UDP on the same IP address and same port.


2.Network Function

It provides the function of connecting / disconnecting the Managed Load Balancer to the logical network and the function of setting the static route.

Item

Contents

Interface

・You can select a logical network created in advance and specify the logical network to connect to.
・You can create up to the maximum number of interfaces specified in the selected service plan.
・You can specify the virtual IP address (VIP) of the interface. It is necessary to specify four reserved IP addresses for each interface regardless of SINGLE configuration plan or HA configuration plan.

Static Route

・You can configure static routes.
・You can set a network address in CIDR format.
・You need to set the next hop IP address.

Note

  • NAT function is not provided.

  • It is necessary to create a logical network and subnet to connect the Managed Load Balancer in advance. See “Restrictions” for notes on network design.

  • It does not connect to the logical network (storage plane). It only connects to the logical network (data plane).

  • You cannot change the virtual IP address and reserved IP address of the interface once connected. If you want to change it, you need to delete and recreate the interface.

  • It takes about 15 to 20 minutes to apply the logical network and interface settings.


3.Instance control function

This function provides the ability to create, change settings, and delete Managed Load Balancer Instance.

Item

Settings

Contents

Create

Managed Load Balancer Instance creation

・You can create a new Managed Load Balancer Instance.
・You can set the name, description, and tags of the Managed Load Balancer at creation time.
 

Service plan selection

・You need to select a service plan when you create it. Determine the service plan from the configuration, bandwidth, and number of ports in advance and select it.
※You cannot change your plan. If you wish to use a different plan, please delete it and create a new one.
 

Select Logical Network

・You can specify the connection destination logical network when creating the load balancer. It is necessary to create a logical network in advance.
 

Interface settings

・You can specify the virtual IP address (VIP) of the interface.
・It is necessary to specify four reserved IP addresses for each interface regardless of SINGLE configuration plan or HA configuration plan.

Deploy

 
・Make various settings provided by the network function and load balancing function, and actually apply them in the Managed Load Balancer instance.
・Billing will start when you deploy.

Setting changes

Apply settings

・You can change various settings provided by the network function and load balancing function.
・The contents set in advance on the screen are regarded as “waiting for application (STAGED)” and are applied by this function.
 

Cancel settings

・You can cancel the setting of “Waiting for application (STAGED)” before it is applied.

Delete

 
・It is a function to delete Managed Load Balancer. Once deleted, the Managed Load Balancer cannot be restored.

Note

  • Reserved IP addresses are IP addresses that are pre-reserved for each interface in order to perform software version upgrades without compromising managed load balancer redundancy. See “Restrictions” for more information on virtual IP addresses and reserved IP addresses.

  • It takes about 15 to 20 minutes to deploy and apply the settings.


4.Management function

Provides functionality for managing a Managed Load Balancer.

Item

Contents

Monitoring function (cooperation with monitoring service)

・It is possible to monitor various information of load balancer resources in cooperation with the cloud / server monitoring service.
・Information (meters) that can be monitored is cloud / server monitoring service manual.
・Please note the following points regarding the meter that obtains the average value for 5 minutes.
――It may not be the average of consecutive acquisition results every minute.
–Multiple meters may be sent with the same time stamp.

User interface

・GUI (SDPF Service Portal) and API can be used for operation.
・For GUI / API, see Management Function
・Please refer to the API reference for the contents that can be operated / acquired by API.

Note

  • Log transfer function to syslog server is not provided.


Description of the behavior of Managed Load Balancer

Instance switching for configuration changes and disaster recovery

  • In this menu, when the customer applies the setting changes or when it is necessary due to fault recovery etc., a new Managed Load Balancer Instance is created in both HA configuration and single configuration, and all settings are synchronized before switching.

  • When switching occurs, communication may be interrupted.

  • After switching, the old Managed Load Balancer Instance will be deleted. Due to the inability to obtain information from an instance that has already been deleted, we may not be able to respond to some inquiries such as the root cause of past events.

sync_instance

Redundancy method of HA configuration (redundancy configuration)

  • In the HA configuration (redundant configuration) plan of this menu, a redundant configuration is set up between two instances created in different groups.

  • For switching the HA configuration, the instance with the highest priority is always set to Active. If a failure occurs in the Primary and a switch occurs, the switch will automatically occur after the Primary is restored.

mLB_HA

Synchronous specifications

  • The following information is synchronized between two instances in an HA configuration or the old instance and the new one when switching them. In addition, it may not be possible to synchronize all information, such as when there is a lot of synchronization information.

No.

Protocol

Synchronous information

1 HTTP / HTTPS

Persistence (source-ip)

2 TCP / UDP

Session, persistence (source-ip)


Service Plan

Service List

This menu offers the following plans.

Plan

Number of interfaces (maximum number)

Configuration

Descriptions

50M_HA_4IF 4

High Availability (Redundant)

The maximum throughput is 50Mbps.

200M_HA_4IF 4

High Availability (Redundant)

The maximum throughput is 200Mbps.

1000M_HA_4IF 4

High Availability (Redundant)

The maximum throughput is 1000Mbps.

3000M_HA_4IF 4

High Availability (Redundant)

The maximum throughput is 3000Mbps.

50M_SINGLE_4IF 4

SINGLE (Non-redundant)

The maximum throughput is 50Mbps.

200M_SINGLE_4IF 4

SINGLE (Non-redundant)

The maximum throughput is 200Mbps.

1000M_SINGLE_4IF 4

SINGLE (Non-redundant)

The maximum throughput is 1000Mbps.

3000M_SINGLE_4IF 4

SINGLE (Non-redundant)

The maximum throughput is 3000Mbps.

Note

  • We recommend an HA configuration (redundant configuration).

  • The bandwidth quality of the plan is best effort, and it does not secure the bandwidth.

  • Maximum throughput is not guaranteed as a guide.


Application Process

Customers who have subscribed to the Smart Data Platform can apply for this menu. The application type, application process, and delivery date are as follows.

Order Types

Application Process

Offering Date

Addition of new Managed Load Balancer

Customers can order via Customer Portal / API

Immediate

Delete Managed Load Balancer

Customers can order via Customer Portal / API

Immediate

Note

  • You cannot change the plan. Please delete / create it.

  • There is no charge at the stage of “Creating” the load balancer. Billing will start when you execute “Deploy”.

  • Please note that billing will continue even if you stop the Managed Load Balancer or disconnect from the logical network. If you want to stop billing, delete Managed Load Balancer.


Application Consideration

  • The maximum number of applications for Managed Load Balancer is as follows.

Limit setting unit

Application items

Maximum Number

Per tenant

Managed Load balancer

16

Per load balancer

Interface

4

  • The following is a list of the upper limit of the number of settings for the load balancing function.

Limit setting unit

Item

Maximum Number

Per load balancer

Health monitor

50
 

listener

50
 

Policy

50
 

Static Route

25
 

Target group

50

Per target group

member

100

Terms and Conditions

Usage Conditions with Other Service Menus

  • There are no specific conditions. Customers may subscribe this menu in combination with all the menus of Enterprise Cloud 2.0.


Minimum Usage Period

  • There is no minimum contract period.


Price

Initial Fee

  • There is no initial cost for this menu.


Monthly Fees

  • The fee for this menu will be charged monthly.

  • Monthly fees will be charged according to the usage time (minutes) of the current month with a monthly upper limit.


Quality of Service

Support Coverage

  • We will support the functions described in the service manual and the settings described in the tutorial.

  • We may not be able to respond to detailed equipment specifications not described in the service manual, or not be able to respond to your design consultation.


Operations

  • The operation of this menu conforms to the operational quality specified as standard in the Smart Data Platform.

  • In addition, as a managed service, we will implement the following operations.

Item

Descriptions

Life Cycle Management of the Products

Implementation of version upgrade

Applies security patches

Apply security patches according to the degree of impact (Same process as version upgrade)

Monitoring / Maintenance

Operation monitoring and failure countermeasure implementation of this device


Maintenance

  • During maintenance work related to Managed Load Balancer, communication may be interrupted in both single configuration and HA configuration.

  • Maintenance will be carried out after prior notice, but the work date and time cannot be adjusted.


Vulnerability

  • We do not inform you about the policy for dealing with all vulnerabilities in this menu. When a vulnerability is discovered, we will post the countermeasures for those that we deem necessary to respond to in the Knowledge center.


Failure definition and notification

  • The definition of failure in this menu is as follows.

Plan

Definition of failure

Notification method

50M_HA_4IF / 200M_HA_4IF / 1000M_HA_4IF / 3000M_HA_4IF

Service interruption for 2 minutes or more

E-mail notification

50M_SINGLE_4IF / 200M_SINGLE_4IF / 1000M_SINGLE_4IF / 3000M_SINGLE_4IF

Service interruption for 10 minutes or more

E-mail notification


About the normality criteria and correspondence of load balancer resources

  • In this menu, the health of the customer’s load balancer instance is monitored by the SNMP response of the instance. If there is no SNMP response for a certain period of time, we will take measures to recover the target resource.

  • While we are monitoring the normality, if the state that we judge that it is not normal continues, we will consider it as “service interruption”. Failure notification will not be given for events that do not meet the definition of failure.

  • The detailed method of confirming normality is not disclosed.

  • You can also check the health of the load balancer from the monitoring status of the detailed information of Managed Load Balancer. If the monitoring status is ACTIVE, the load balancer is working properly.


SLA

  • This menu is not covered by the SLA.


Restrictions

About load balancer specification change

  • Since this menu uses load balancer software internally, the specifications may change due to software version changes or product changes.


About configuration and design

  • It is not possible to take a one-arm configuration. See the quick start guide in the tutorial for a configuration example.

  • You cannot set the IP address of the same segment as the virtual IP address (VIP) of the interface of the load balancer to the IP address of the listener. Please set the IP address of the listener prepared in another segment.

  • It is the customer’s responsibility to design the IP address in the logical network to which this menu is connected.

  • It is recommended to set a wide range of addresses for subnets belonging to the logical network to have the extensibility of resource expansion

  • It does not connect to the logical network (storage plane). It only connects to the logical network (data plane).

  • Reserved IP addresses are IP addresses that are pre-reserved for each interface in order to perform software version upgrades without compromising managed load balancer redundancy.

  • The address specified by the virtual IP address (VIP) cannot be used as the reserved IP address.

  • It is not possible to register an address in the address band of ISP shared address (100.64.0.0/10) as the virtual IP address (VIP) and reserved IP address.

  • You cannot use the IP address (virtual IP address (VIP), reserved IP address) assigned to this menu for the physical interface of your other device, permitted IP address, etc.

  • You cannot change the virtual IP address and reserved IP address of the interface once connected. If you want to change it, you need to delete and recreate the interface.

  • It is not possible to listen to both of TCP and UDP on the same IP address and same port.


Setting upper limit for each function

  • There is an upper limit to the number of settings for the load balancing function, so please refer to the upper limit list in “Application Consideration”.

  • Only one health monitor / listener / target group can be specified in one policy.

  • Interface limits vary depending on the plan.


performance

  • This menu is provided by installing Managed Load Balancer on Server Instance. Therefore, the number of sessions is subject to the constraints of Server Instance on which Managed Load Balancer is installed. For details, see Instance restrictions from the detailed information in the Server Instance menu.

  • The maximum throughput of each plan is just a guide. We do not guarantee the numerical value.

  • We have released The performance measurement results of Managed Load Balancer. The maximum value of each performance item was measured, and not all the maximum values of each performance item were measured at the same time. Please note that this verification result is a reference value only and does not guarantee the performance.


Others

  • Each function and log provided by this menu does not ensure integrity, accuracy and compatibility for customer’s purpose of use.