Scrutinised and well-considered security processes are a critical part of delivering a successful product to our customers.
Network level security consists of three main components:
- DDoS mitigation
- VLAN reverse path forwarding protection
- Juniper firewall rules at the network edge and core
A DDoS detection and mitigation system is deployed at our data-centre. DDoS attack traffic is diverted to a filter/scrubbing server that can distinguish between valid and malicious traffic. Malicious traffic is scrubbed off while valid traffic is re-injected into the network. The victim IP is not affected during the DDoS attack. DDoS detection and mitigation is fully automated and traffic diversion occurs automatically.
Small DDoS attacks are scrubbed locally in the data-centre by the mitigation system. For larger attacks, traffic is diverted to an international DDoS mitigation provider which then sends the clear traffic on to South Africa.
Reverse path forwarding protection is enabled for all VLANs in our data centre. This policy ensures that only the subnets allocated to a VLAN can generate traffic for that VLAN. This helps to mitigate two kinds of malicious traffic:
- Source-spoofed traffic where a host is sending out traffic for subnets that do not belong to the VLAN.
- Inter-VLAN subnet spoofing, where a host in one VLAN uses IP addresses from another VLAN using source-spoofing.
Firewall rules on the data centre network edge and at the core are used to protect the network in a number of ways:
- Rate-limiting of certain protocols to protect the network infrastructure.
- Blocking of certain protocols and destination IP addresses to protect Pixel Forest operational systems.
- Restricting access to certain hosts and protocols to defined lists of source addresses.
- Blocking of abusive IP addresses and hosts.
All servers are monitored 24/7 for all critical services and hardware health. Our reactive system administrators react to monitoring alerts as they are identified and escalate issues to the data centre.
Servers are designed to provide redundancy and reliability, including multi-core, multi-CPU systems, ECC (Error-Correcting Code) memory modules to detect and correct data corruption in real time and enterprise grade storage that includes hard disk and solid state drives.
All data is stored on dedicated, robust RAID storage arrays providing data redundancy and integrity.
All Pixel Forest Servers are automatically backed up in the early hours of the morning. The backup includes all critical data required for disaster recovery.
Backups are made of the user’s home directory as well as databases. The user’s home directory will include site content, web logs and any mail that was on the server at the time that backup was completed.
Customers can restore up to the previous 2 weeks of backup data via the konsoleH control panel. Please note that Pixel Forest does not guarantee backups. If you have critical data which you cannot afford to lose in the event of a disaster, keep a copy of your data locally (or at an alternate location) as well.
Logs (such as FTP, web server and mail logs) are normally kept for 60 days.
Due to the large scale of our Web hosting, our backup and restore process is effectively tested on a daily basis.
All servers (which are Linux based) run Clam anti-virus which is updated as new virus definitions are released. Servers are scanned daily.
All customer passwords are stored in a one-way encrypted format. Pixel Forest is not able to retrieve any passwords. Due to the broad technology implementation across our hosting software and platform, we employ a number of different password hashing algorithms e.g. bcrypt, sha-512. We implement industry standard practices for mitigating various password cracking methods e.g:
- Password salts to mitigate rainbow attacks
- Multiple password hashing rounds (key stretching) to massively draw out brute force attacks
SSL is used for POP, IMAP and SMTP protocols for email, resulting in data encryption between our server and customers’ mail programmes.
The use of strong passwords is enforced when creating or editing mailboxes via the mail admin tool.
The following measures are used to mitigate spam and malware:
- Anti-virus and anti-spam scanning occurs on all inbound and outbound email.
- Common malicious file extensions are blocked for both inbound and outbound email.
- Known malicious IP addresses are blocked by our firewall for incoming email.
While we care for the hosting infrastructure including the network and servers, it is our customers responsibility to keep their data and hosting account secure.
- Use secure passwords and store them safely
- Ensure sufficient security for your web applications
- Ensure that CMS’ and plugins are always kept up-to-date