Loadbalancer.org are one of a number of vendors that pride themselves on offering affordable load balancing appliances that work. It is the likes of such companies that have collectively driven down the price of these solutions, making load balancing appliances available to companies who previously would not have been in a position to consider such investments. Kemp Technologies are a similar company who's primary marketing drive centre's around 'value for money'. It is because of this glaring similarity that I decided to compare SSL performance capabilities, focusing on the entry-level appliance on offer from each vendor. Specification comparisons were taken, and subsequent performance tests examined whether performance levels met that of the stated specification. Results proved extremely interesting!
The Enterprise R16 can handle 7 million+ simultaneous load balanced connections using 1 GB RAM and offering layer 4 and 7 load balancing. With full layer 7 proxy support including cookie insertion/inspection and, of particular importance to this entry, the R16 claims SSL termination at 700 TPS. An Enterprise R16 will set you back £1,495. Kemp's budget appliance, the LM1500, offers layer 4 and 7 server load balancing with up to 256 virtual and 1000 real servers. This solution will give you up to 100 Mbps throughput and set you back £1,300. This appliance is capable of layer 7 content switching and claims SSL acceleration up to 100 TPS. It's all well and good comparing the specifications but for a truly valid comparison I have performance tested each appliance. To ensure validity and reliability of results I tested 300 requests at a concurrency level of 35 on both machines. More specifically, for the Kemp I used the command
$ ab -n 300 -c 35 http://192.168.2.184:80/
while for the R16 the command was
$ ab -n 300 -c 35 http://192.168.2.21:443/
The subsequent results proved extremely interesting, and somewhat surprising! Whilst the Kemp LM1500 claimed up to 100 TPS, the appliance actually fell far short of this managing a mere 58 TPS. Conversely, the R16 claimed a vastly superior 700 TPS and actually significantly exceeded this managing 815 TPS.
The mean time per request (across all concurrent requests) for the LM1500 was 17.221 (ms) while for the R16 it was significantly faster at 1.228 (ms).
When looking at the transfer rate the Loadbalancer.org appliance continued to out-perform the Kemp receiving 236.22 Kbytes/sec as opposed to a mere 18.58 Kbytes/sec with the LoadMaster.
Furthermore, the Enterprise R16 served 100% of the requests in 97 milliseconds, whereas the Kemp could not serve 50% of the same number of requests in this time (50% in 102 milliseconds). And with the maximum connection time taking a massively longer 5113 milliseconds it is amazing that these two appliances are in the same price bracket.
To view the full performance test results click here. These results are somewhat shocking when you consider that both solutions are priced within a couple of hundred pounds of each other. For SMB's to invest in a load balancing appliance these results would show only one option, in terms of SSL acceleration. The Loadbalancer.org appliance significantly out-performed its competitor, the Kemp LM1500 and would without doubt be the wise choice for such an investment.