Limit all CPU, IO, memory resources, numbers of processes, and concurrent connections per each user
A single site can easily bring your server to a halt by consuming all your CPU, memory, and IO resources. Our proprietary Lightweight Virtual Environment (LVE) technology prevents that by allowing hosts to set up individual resource limits. This ensures that a tenant can never use more resources than he or she is given.
LVE is a kernel-level technology developed by the CloudLinux team. It integrates at the server, PAM (Pluggable Authentication Modules), and database levels to prevent any kind of abuse while maintaining the lowest overhead possible. The technology has roots in common with container-based virtualization.
The goal of LVE is to ensure that no single website can bring down your server.
LVE benefits include:
Ensure Utmost Stability
By limiting resources, you ensure that a tenant never uses more than he is allowed, and therefore never slows down or brings down your entire server or causes issues for all other tenants on the server.
Increase Revenue
LVE Manager acts as a powerful sales tool – you can set limits on a per-package basis and enjoy this feature as a new upsell opportunity, creating packages with more CPU and Memory and selling them to those customers who use more and need it. The best part is that your end users can see their own usage on the account, which makes upsell much smoother.
Reduce Churn Rate
Because the resources of each tenant are limited, your servers won’t go down or slow down, and this will minimize inquiries for performance issues – the primary cause of customer dissatisfaction.
LVE Manager
LVE Manager allows you to maintain fine-tuned control over your resources, including CPU, IO, memory, inodes, numbers of processes, and connections, that any single account can use. It is lightweight and transparent. Now you can limit abusers while allowing good customers to use what they need.
With LVE Manager, you can:
- Limit resources per single account – for end-users and for resellers
- Allow resellers to limit resources for their end-users
- Create and apply default packages
- View usage history per account
- Identify abusers and take corrective actions
- Identify top users and upsell to higher-end plans
Memory
Memory limits control the amount of memory each customer can use. CloudLinux is able to identify, in real time, the amount of memory actually used by an end customer’s processes. Physical memory limits are especially effective in preventing out of memory (OOM) issues and customers’ ballooning memory usage, which destroys caches and causes server overload.
IO
IO limits restrict the data throughput for the customer. They are measured in KB/s. When the limit is reached, the processes are throttled (put to sleep). Because IO is one of the scarcest resources in shared hosting, the ability to put an upper limit on customer use is vital.
CPU
CPU limits establish the maximum amount of CPU resources that an account can use. When a user hits the CPU limit, processes within that limit are slowed down. CPU limits are crucial in preventing CPU usage spikes, which can often make servers slow and unresponsive.
Number of Processes
Number of processes limits control the total number of processes within LVE. Once the limit is reached, no new process can be created until another one has finished. This effectively prevents fork bombs and similar DoS attacks.
Entry Processes
Entry processes limits control the number of entries into LVE. The best way to think about this type of limit is as the number of web scripts that can be executed in parallel by visitors to a site. These limits are important to preventing single sites from hogging all Apache slots, thus causing Apache to be unresponsive.
Inode
An inode is a data structure on a file system that is used to keep information about a file or a folder. The number of inodes indicates the number of files and folders an account has. Inodes limits work on the level of disk quota.
taken from