|Before introducing SIOS Coati||After introducing SIOS Coati|
|AWS Auto Scaling enables automatic scaling up/down
of web servers according to the accesses volume
|With SIOS Coati, automatic operations of Amazon EC2,
which is out of scope of AWS, at the application level
|The company outsources manned operation service
during the daytime, but the company employees
should take care of the operations during the night
|Leaving the night operations to SIOS Coati gives operation personnel peace of mind|
|It takes time every day to address system troubles||SIOS Coati provides maintenance-free and care-free
operations at low cost so that the company can focus
on providing the high quality content
Hobby Japan is a long-established publishing company that celebrated its 50th anniversary, which develops diverse businesses related to hobbies (e.g. anime, figures, games, etc.). "Hobby Japan" Magazine, bearing the company's name, has recorded 600th issue. Also the company publishes magazines including "Monthly Arms Magazine" and "Card Game," and young-adult books. In addition, the company also conducts businesses such as selling figures and character goods, game planning, and importing and selling analog games. Fukabori Yuichiro from Public Relations, Business Administration Department of Hobby Japan introduces the company as "a general trading company of hobbies that handles a wide range of things related to hobbies."
Hobby Japan, where hobbyists who know everything about various hobbies and want to entertain people gathered, implemented SIOS Technology’s "SIOS Coati" as a cloud operation management tool in April 2019. For Hobby Japan employees, operations of the website is mission-critical because it is a source of information and important business resource: therefore, being free from the fear of server down enables employees to focus on their essential creative work.
New problems occurred even after the transition to AWS aimed at stable operations
Looking back the history of Hobby Japan's information system, it originally ran a web server by renting dedicated servers from a hosting server company. However, rented servers became obsoleted and it was also burdensome to take care of updating. Furthermore, it took time to scale servers as the business expanded. Under such circumstances, a major incident occurred: the HDD data used by the character business web server was gone.
"We were able to recover the data finally, but since no security measures or data backup had been taken, it took a lot of time to recover. Therefore, we wanted to switch to a safer system. In addition, because cloud services were getting popular around that time, we started to consider using a cloud service," said Mr. Fukabori.
Then, Mr. Fukabori actually experienced a cloud service. "I was so surprised that the server started up with just one button in 10 minutes. That's why I thought that if we could use this," Mr. Fukabori recalls his experience. However, at that time he could not find any company to be a partner, so Mr. Fukabori himself collected information from the Internet and sought a transition to a cloud service.
Mr. Fukabori, who currently is a member of Public Relations, had been in the company's web media department before and since then he has been taking care of company-wide web server systems. "Our company has a tradition where employees are in charge of multiple functions, and various tasks are progressing simultaneously across the company. Moreover, since there is no IT department within the company, I am in charge of operations of all servers except those which managed by each department," said Mr. Fukabori. It is easy to imagine that he is in a difficult situation.
With this background, Mr. Fukabori talked to the upper management to obtain a permission for migrating to a cloud service in order to realize stable operations, and finally he got an approval. They migrates to Amazon Web Services (AWS) in 2013, and they should have been freed from various problems of the rental servers.
However, luckily or unluckily, access to the website of Hobby Japan steadily increased at the same time. That's good for business, but for Mr. Fukabori, who manages AWS together with other responsibilities, was a heavy burden.
"Sometimes volume of access to the website becomes quite high due to various factors, such as the publication of comics on the website and an animation based on the cartoons of Hobby Japan being broadcasted on television, etc. As a result, the server was often down due to the overload. I was blamed by the business departments," said Mr. Fukabori.
Server down due to the access overload sometimes occurred in the middle of the night and Mr. Fukabori got a call while he was asleep. Finally he “was taken aback just by receiving incoming calls,” added Mr. Fukabori. It is understandable that he wanted to bring back a “normal lifestyle.”
He thought “I can’t handle this by myself” and asked AWS for help. AWS introduced him ForgeVision,Inc, a company which provides cloud service integration.
Implement SIOS Coati in order to realize more stable operations
Before talking to ForgeVision, Mr. Fukabori monitored AWS EC2 instances with AWS Cloud Watch. But he was
not able to do anything even when the access overload was expected such as a release day of Monthly
Hobby Japan. Insted, he only got a notification at the time of server down. Moreover, ”when our content
was introduced on a popular video site or our figures were buzzed on Twitter, it caused unexpected
increase in access to the server, and the server was down," he said. He had system troubles about 10
times a month and he was busy addressing them.
Based on such circumstances, ForgeVision proposed three solutions and implemented them in August 2018. The first solution was "implementing an auto-scaling to EC2 instances" that enables scalable response when access overload occurred. The second was the “introduction of a visualization tool” that allowed operators to check access status and resource utilization. The third is "reducing operation loads" by introducing AWS operations management service.
Auto Scaling realized stable server operations by automatically scaling a minimum of one server and 2~3 servers for instantaneous access overload. "We did not have any trouble on the summer animation broadcasting day in 2018 when many users access our server simultaneously thanks to auto scaling," said Mr. Fukabori. The number of troubles he had in a month was reduced from 10 times to 1 or 2 times, and he could return to a “normal lifestyle.”
However, not all the system troubles could be taken care of. That's because ForgeVision's AWS operations and management service delivers system failure notifications 24/7, but failure is addressed by people only between 10am to 18pm. Therefore, troubles occurred after 6pm is left as it is until the next day. For Hobby Japan, because the access to their website tends to be increased at night, this has been a bit disappointing. When such issues came to light, ForgeVision proposed a cloud operation management tool called "SIOS Coati."
Focus on exploring the interesting content with maintenance-free system
The SIOS Coati constantly monitors EC2 instances and has the ability to automatically restart and recover if instances or applications fail. Since most failures can be resolved by restarting the instance or application, this feature provides peace of mind even at night when people cannot address a failure. Mr. Fukabori decided to implement SIOS Coati at Hobby Japan, expecting it can bring the sense of security in the operations management.
Hobby Japan started using SIOS Coati from April 2019. Though no failure which SIOS Coati needs to address occurred during the first two months, Mr. Fukabori can always have a peace of mind with a fail-safe mechanism that “automatically recovers the system” even if a failure occurs.
"Of course, all the troubles due to setting errors in addition to the over-access cannot be prevented. But I think that the introduction of SIOS Coati is very good for my peace of mind. If we have a system which automatically recover immediately in case of failure, operators will not be called every time a failure occurs," said Mr. Fukabori.
In the first place, it is not desirable for the persons in charge of operations and management of the information system to spend much time on “babysitting” the system. "I don't want to care about the system or server as much as possible. I would like to completely outsource it if we could, but it would cost more," said Mr. Fukabori. “Capability to enable maintenance-free, care-free environment at low cost is the SIOS Coati’s advantage.”
He also added; “since Hobby Japan is a company to produce content, employees want to pursue the interesting content. We produce high quality content, but it will go in vain if we cannot provide it at the right timing due to the server down. Therefore, we need a system environment where we even don’t have to manage. In that sense, I feel that we are one step closer to an environment where I can stay calm even when too many people access to the popular content.”
At the moment, SIOS Coati only monitors some of EC2 instances used by Hobby Japan. Currently they are verifying the effect with the largest instance and are also considering increasing the monitoring targets in future. Migration to AWS, use of the ForgeVision operation management service, and introduction of automatic recovery capability with SIOS Coati. With these steps, it seems that employees of Hobby Japan are steadily approaching an environment where they can provide the best content with peace of mind.