matomo

Transparency Report: June 2015

David Wibergh, about Status & transparency reports

This is the ninth summary we publish. The summary shows the amount of traffic passing through our servers, uptime and peak traffic. We also report how many customers we have and plans for the near future.

OVPN had two planned maintenances in June. We were also affected by the global disruption of the Internet that occurred on June 12.

Summary

Our VPN servers maintained an uptime of at least 99.5% in June.

Traffic decreased by 45 TB in June from the previous month, which means that the total volume of traffic was 365 TB.

In June, OVPN received an additional 300 customers, bringing the total number to 2900. We are also preparing to grow very much in the autumn.

Within two months we release our own client for Windows, Linux & Mac OS X. We work hard on the client and are very pleased with the results. The client is simplifying the process of using OVPN enormously because no special settings need to be made. In addition to simplifying the use of OVPN, the client also has a number of functions that strengthen security and anonymity even more for our customers.

We are also preparing for an international expansion, which is done by renting rack cabinets in other countries from Europe. These rack cabinets and the servers placed in them will have exactly the same physical security as our data centers in Sweden. Negotiations are underway with several data centers and internet service providers to find the best solution for our needs. We have also started the process of translating the website into several languages and improving all content. In connection with this, we rewrite the entire codebase of the website in Laravel 5.1.

As mentioned earlier, the infrastructure in Stockholm will also be upgraded, which is done by acquiring several servers and a new switch. VPN05 will soon also be updated so that the server can assign both Swedish and US IP addresses.

We recently released our disk image for OVPNbox as open source. Instructions and the disk image can be found in our FAQ. If you don't have a computer that you can use as a box, you can buy it from us. However, there are only nine boxes left in the warehouse - after that it is completely sold out! If you are interested in renting or buying a box, you should take the opportunity as it is the last boxes we will sell this year.

Inthe future, we are working to load balance the website to ensure that a downtime on a web/sql server does not cause any problems with our API. The solution to ensure a high level of reliability is structured in two separate steps:

  1. Website and MariaDB servers will soon be behind a load balancer. We also operate another server, which means that web & sql runs on three separate servers in three different data centers. This means that all three servers need to be down in order for our VPN servers not to be able to authenticate clients.

  2. If the worst happens and all servers are down, all VPN servers will allow anyone to connect while we cannot authenticate clients. When web & sql servers become available again, active clients will be reauthenticated and 'unauthorized' will be logged out. This will work for all services except public IP because the servers do not know which IP address to assign to the client.

Below is the raw data and statistics for all servers in June.

VPN01 - Malmö

Traffic (Mbit/s)

VPN01 - Summary of traffic peaks in June

Analysis

VPN01 encrypted and anonymized a total of 53,827 GB (52.56 TB) traffic. That's an increase of 6,673 GB from the previous month.

Peak traffic was 635 Mbps, which is 44.6% of maximum capacity.

95th percentile was 327 Mbps, meaning that 95 percent of the time VPN01 was below 22.9% of maximum capacity.

VPN01 had an uptime above 99.5% during June.

VPN02 - Stockholm

Traffic (Mbit/s)

VPN02 - Summary of traffic peaks in June

Analysis

VPN02 encrypted and anonymized a total of 71,107 GB (69.44 TB) traffic. That's a decrease of 10,035 GB from the previous month.

Peak traffic was 865 Mbps which is 60.8% of maximum capacity.

95th percentile was 429 Mbps, which means that 95 percent of the time VPN02 was below 30.1% of maximum capacity.

VPN02 had an uptime above 99.5% during June.

VPN03 - Stockholm

Traffic (Mbit/s)

VPN03 - Summary of traffic peaks in June

Analysis

VPN03 encrypted and anonymized a total of 74,374 GB (72.63 TB) traffic. That's a decrease of 11,760 GB from the previous month.

Peak traffic was 800 Mbps, which is 56.2% of maximum capacity.

95th percentile was 414 Mbps, which means that 95 percent of the time VPN03 was below 29.1% of maximum capacity.

VPN03 had an uptime above 99.5% during June.

VPN04 - Stockholm

Traffic (Mbit/s)

VPN04 - Summary of traffic peaks in June

Analysis

VPN04 encrypted and anonymized a total of 58,612 GB (57.24 TB) traffic. That's a decrease of 15,930 GB from the previous month.

Peak traffic peak was 678 Mbps, which is 47.6% of maximum capacity.

95th percentile was 324 Mbps, which means that 95 percent of the time VPN04 was below 22.7% of maximum capacity.

VPN04 had an uptime above 99.5% during June.

VPN05 - Stockholm

Traffic (Mbit/s)

VPN05 - Summary of traffic peaks in June

Analysis

VPN05 encrypted and anonymized a total of 10 367 GB (10.12 TB) traffic. That's an increase of 1,040 GB from the previous month.

Peak traffic was 373 Mbps, which is 26.2% of maximum capacity.

95th percentile was at 109 Mbps, which means that 95 percent of the time, VPN05 was below 7.67% of maximum capacity.

VPN05 had an uptime above 99.5% during June.

VPN06 - Malmö

Traffic (Mbit/s)

VPN06 - Summary of traffic peaks in June

Analysis

VPN06 encrypted and anonymized a total of 34,776 GB (33.96 TB) traffic. That's an increase of 5,322 GB from the previous month.

Peak traffic was 713 Mbps, which is 50.1% of the maximum capacity.

95th percentile was 258 Mbps, which means that 95 percent of the time VPN06 was below 18.1% of maximum capacity.

VPN06 had an uptime above 99.5% in June.

VPN07 - Malmö

Traffic (Mbit/s)

VPN07 - Summary of traffic peaks in June

Analysis

VPN07 encrypted and anonymized a total of 33 656 GB (32.87 TB) traffic. That's a decrease of 11,023 GB from the previous month.

Peak traffic was 608 Mbps, which is 42.6% of maximum capacity.

95th percentile was 249 Mbps, meaning that 95 percent of the time VPN07 was below 17.50% of maximum capacity.

VPN07 had an uptime above 99.5% during June.

VPN08 - Malmö

Traffic (Mbit/s)

VPN08 - Summary of traffic peaks in June

Analysis

VPN08 encrypted and anonymized a total of 37,412 GB (36.54 TB) traffic. This is a decrease of 9,509 from the previous month.

Peak traffic was 493 Mbps, which is 34.7% of maximum capacity.

95th percentile was 237 Mbps, meaning that 95 percent of the time VPN08 was below 16.6% of maximum capacity.

VPN08 had an uptime above 99.5% during June.

WEB01 - Falun

Uptime (%)

WEB01 - Summary of uptime in June

WEB01 had an uptime of 99.76% in June, resulting in a total downtime of 101 minutes.

Downtime was largely due to the global outage, which affected access to the server. There was also a problem with the web server in connection with this, which prolonged the outage.

SQL01 - Falun

Uptime (%)

SQL01 - Summary of uptime in June

SQL01 had an uptime of 99.94% in June, resulting in a total downtime of 27 minutes.

Down time was largely due to the global outage, which affected access to the server.

SQL02 - Falkenberg

Uptime (%)

SQL02 - Summary of uptime in June

SQL02 had an uptime of 98.60% in June, resulting in a total downtime of 585 minutes.

The downtime occurred because the data center in which SQL02 is located had a long malfunction that affected the operation of the server.

David Wibergh