As a system administrator for a number of years and prior web hosting business owner, it pains me to see so many negative comments.
I do not really want to get involved but as a guy on a REALLY strict budget I avoid 99% of the sprint runs and special offers...it's all about the benjamins and the mad rush to get the product. I just choose not to fight it.
Almost all of you would be surprised that it does not take that many people hitting RELOAD to overwhelm a website. Spyderco is not likely to lease bandwidth costing many tens of thousands of dollars for a one day seconds sale.
Almost all of you would be surprised that it does not take that many people hitting RELOAD to overwhelm a website. Spyderco is not likely to lease bandwidth costing many tens of thousands of dollars for a one day seconds sale.
Fair enough. From a sys admin's viewpoint, what are possible solutions, while staying within existing bandwidth constraints?
Steel novice who self-identifies as a steel expert. Proud M.N.O.S.D. member 0003. Spydie Steels: 4V, 15V, 20CV, AEB-L, AUS6, Cru-Wear, HAP40, K294, K390, M4, Magnacut, S110V, S30V, S35VN, S45VN, SPY27, SRS13, T15, VG10, XHP, ZWear, ZDP189
Almost all of you would be surprised that it does not take that many people hitting RELOAD to overwhelm a website. Spyderco is not likely to lease bandwidth costing many tens of thousands of dollars for a one day seconds sale.
Fair enough. From a sys admin's viewpoint, what are possible solutions, while staying within existing bandwidth constraints?
The continuous reloading of web pages on a domain is called a "DoS' or "DDoS" which stands for denial of service or distributed denial of service.
99% of the systems on the Internet do not have mechanisms to deal with these issues. CloudFlare is one company that offers solution to DoS and DDoS, and hackers the world over use DoS to make services unreliable, unavailable and to hide attacks on systems.
If I had a solution, I would be building it right now and I'd make a billion. There is no known solution to these crises other than routing requests made to a server to /dev/null
I'm just guessing, but if there were 2500+ shoppers on the site and 1000 were adding items to their shopping carts and constantly reloading their shopping carts BOOM, that is all that it takes. One server that can handle that small number of shoppers for 4hrs would cost more than $50,000. Then for the rest of the year that server sits there devaluing by 33% every year....doing almost NOTHING.
Seen a few great 2-knife boxes on YouTube. Last one was a Manix XL 4V and a S90V Yojimbo. Now that’s a gift from Spyderco.
Yeah that was a joke if you keep watching lol. None of the two knife boxes have been as good as the three knife boxes from what I’ve seen and my experience, but still cool imo
Almost all of you would be surprised that it does not take that many people hitting RELOAD to overwhelm a website. Spyderco is not likely to lease bandwidth costing many tens of thousands of dollars for a one day seconds sale.
Fair enough. From a sys admin's viewpoint, what are possible solutions, while staying within existing bandwidth constraints?
The continuous reloading of web pages on a domain is called a "DoS' or "DDoS" which stands for denial of service or distributed denial of service.
99% of the systems on the Internet do not have mechanisms to deal with these issues. CloudFlare is one company that offers solution to DoS and DDoS, and hackers the world over use DoS to make services unreliable, unavailable and to hide attacks on systems.
If I had a solution, I would be building it right now and I'd make a billion. There is no known solution to these crises other than routing requests made to a server to /dev/null
I'm just guessing, but if there were 2500+ shoppers on the site and 1000 were adding items to their shopping carts and constantly reloading their shopping carts BOOM, that is all that it takes. One server that can handle that small number of shoppers for 4hrs would cost more than $50,000. Then for the rest of the year that server sits there devaluing by 33% every year....doing almost NOTHING.
Thanks for this explanation.
I had wondered what the effect of ‘x’ amount of customers having two browsers constantly refreshing would have on the site.
And now I know.
'The future is already here;it's just not evenly distributed'
William Gibson
...If I had a solution, I would be building it right now and I'd make a billion. There is no known solution to these crises other than routing requests made to a server to /dev/null ...
Are there other solutions besides technical, server-based solutions? For example, does a medium-sized company want 2500+ shoppers hitting their servers all at once? Is that volume of shoppers even desirable, if the infrastructure doesn't exist to support it? Do unsuccessful, frustrated customers try harder and spend more at the next opportunity? Do they decide to pay full retail if they can't buy a second? Or do they become sullen, abandon ship, nurse a grudge, and shop a different brand? I've never marketed a top-flight knife company so I've no idea; maybe scarcity is the strategy. Again, I don't know, but it's an interesting case study to observe first-hand. You do make the case for why "just adding servers" is not a logical solution.
Steel novice who self-identifies as a steel expert. Proud M.N.O.S.D. member 0003. Spydie Steels: 4V, 15V, 20CV, AEB-L, AUS6, Cru-Wear, HAP40, K294, K390, M4, Magnacut, S110V, S30V, S35VN, S45VN, SPY27, SRS13, T15, VG10, XHP, ZWear, ZDP189
...If I had a solution, I would be building it right now and I'd make a billion. There is no known solution to these crises other than routing requests made to a server to /dev/null ...
Are there other solutions besides technical, server-based solutions? For example, does a medium-sized company want 2500+ shoppers hitting their servers all at once? Is that volume of shoppers even desirable, if the infrastructure doesn't exist to support it? Do unsuccessful, frustrated customers try harder and spend more at the next opportunity? Do they decide to pay full retail if they can't buy a second? Or do they become sullen, abandon ship, nurse a grudge, and shop a different brand? I've never marketed a top-flight knife company so I've no idea; maybe scarcity is the strategy. Again, I don't know, but it's an interesting case study to observe first-hand. You do make the case for why "just adding servers" is not a logical solution.
I have not poked around to see if Spyderco has their site hosted or they handle everything internally. I am not that savvy anymore, I've been out of the loop for a while as far as network operations and web hosting. That was 20+ years ago, and it is not as simple as reading a Whois (but it can be) and I do not run Wireshark for packet capture to read all of the data services involved.
I am not an expert in this area but I'll tell you what I know. I have a BS in MIS and completed all of the training for the Cisco CCNA Internetworking Certificate. I used to build websites, then got into web hosting, then degreed and left for Silicon Valley. Worked in Network Operations for Keynote Systems, which has since been purchased, then worked as a system administrator at a biotech during the complete redevelopment of their IT infrastructure. I was sort of point man on the entire co for 2+ years. But I'm a newb in many ways. Without my boss at the biotech, who worked all of 3 to 4hrs a day I would have had to bootstrap up to speed much faster. Technical support is vast.
There are a host of solutions from: increasing the bandwidth to the retail web server; server tuning which includes everything from maxing out RAM and CPUs(assuming there are multiple slots), to changing the cache size and number of client requests the server can handle. Some of these things you order to have installed and other things you may hire a consulting team to handle, assuming Spyderco needs the assistance.
Spyderco is a knife manufacturer. They do have web security as revealed by Vivi's blocking on the site during the order process. They are not a web hosting company so I assume they farm out a lot of the work to mid to large sized companies.
If Spyderco had 2500+ people at all hours of the day shopping on their site they would be more than happy to buy the extra resources whatever they may be. Infrastructure and support pays for itself if you are selling. Spyderco systems are not unlike Amazon.com during sales. They get hammered. Depending on how often you have sales is what drives the addition of extra capacity.
The possible answer would be to switch to a scalable ecommerce platform, and an overhaul of the website UX. Scalability would allow the site to increase or decrease its capability to serve a number of customers based on demand. The question is if Spyderco would be willing to invest in this.
...If I had a solution, I would be building it right now and I'd make a billion. There is no known solution to these crises other than routing requests made to a server to /dev/null ...
Are there other solutions besides technical, server-based solutions? For example, does a medium-sized company want 2500+ shoppers hitting their servers all at once? Is that volume of shoppers even desirable, if the infrastructure doesn't exist to support it? Do unsuccessful, frustrated customers try harder and spend more at the next opportunity? Do they decide to pay full retail if they can't buy a second? Or do they become sullen, abandon ship, nurse a grudge, and shop a different brand? I've never marketed a top-flight knife company so I've no idea; maybe scarcity is the strategy. Again, I don't know, but it's an interesting case study to observe first-hand. You do make the case for why "just adding servers" is not a logical solution.
While it seems unbelievable, the seconds sale is tiny fraction of their business. Even the amount of people who own a Spyderco on this forum of any knife forum is a minority. I’ve seen tons of people commenting here and other Spyderco forums that they didn’t even know there was such thing as seconds lol. Most companies don’t even offer them. I would say the strategy is just getting rid of product taking up space while also doing a little bit of fan service. Sure, some people feel shafted and get bitter, but I don’t think most do. I didn’t get anything the first year they did it online and I was okay with that, kept buying their knives. Did a little better this year and I’m happy with it. I didn’t get any insane deals in my mystery boxes, but they were cool enough. I will buy them every year they do it, I just hope I have more money next year. Some of the scalpers suck, but that’s life. I think the slow shaky server actually slows them down some lol and gives the rest of us a chance. They could adopt a queuing system, but really they need a new website in general. It’s very dated even aside from the sale.
2 knife Mystery box
Got a PITS Sprint Run, really nice knife. My first foreign made Spyderco.
Second one is a byrd of some sort as I know nothing of the byrd line. If someone could please inform me on it, I would greatly appreciate it.
20230728_173624.jpg
The possible answer would be to switch to a scalable ecommerce platform, and an overhaul of the website UX. Scalability would allow the site to increase or decrease its capability to serve a number of customers based on demand. The question is if Spyderco would be willing to invest in this.
I am hoping it does not have to be a big server cost for them but instead a hybrid of what they have done in the past.
The surprise November sale ran smooth, I did get one knife but also missed one or two as I had no idea the sale was coming and stuff sold out. On the flip side something like this is extremely easy to miss and therefore not ideal.
This second sale was well advertised in advance, if you like Spyderco at all you knew about it. I was on it but could get nothing as the advertising drew way too many people killing the servers.
Maybe the next sale they can skip social media, just put it up in the forum, it will leak out anyway. Maybe not advertise it a week ahead but 24-48 hours too.
Not sure if this would help but it might be a happy medium of notification while not bringing the whole world down on the servers.
The answer might be developing a virtual waiting room that is completely separate from Spyderco.com. When the sale starts, you are placed in queue in the waiting room with an ETA. Once it’s your turn, you get redirected to the seconds page. You then have 10 minutes to browse and place an order. This would eliminate the site getting hammered and give everyone a fair chance at ordering.
The answer might be developing a virtual waiting room that is completely separate from Spyderco.com. When the sale starts, you are placed in queue in the waiting room with an ETA. Once it’s your turn, you get redirected to the seconds page. You then have 10 minutes to browse and place an order. This would eliminate the site getting hammered and give everyone a fair chance at ordering.
I am not a website or server expert but it seems an intestesting idea.
The answer might be developing a virtual waiting room that is completely separate from Spyderco.com. When the sale starts, you are placed in queue in the waiting room with an ETA. Once it’s your turn, you get redirected to the seconds page. You then have 10 minutes to browse and place an order. This would eliminate the site getting hammered and give everyone a fair chance at ordering.
Could be a great idea, but you would just need to prove they are all human and not bots. Maybe have 50 people shopping, one checks out and comes in.
On the flip side it would make one order per person and limit the traffic. People would not have to panic buy what they really want then come back for 2nd/3rd orders when they can look a bit.
If we agree that Spyderco is a company that primarily likes to sell its products through its dealer network what is to prevent Spyderco from reaching an agreement with some of their larger online retailers to present the seconds for sale.
The knives would still be as is, notched, and clearly understood to be seconds, so there would be no question of tainting the first run product stream..
If we agree that Spyderco is a company that primarily likes to sell its products through its dealer network what is to prevent Spyderco from reaching an agreement with some of their larger online retailers to present the seconds for sale.
The knives would still be as is, notched, and clearly understood to be seconds, so there would be no question of tainting the first run product stream..
They could drop ship directly from Spyderco.
I think this approach has potential.
For discussion's sake, I agree that this has potential and will eliminate the need for overhauling the ecommerce platform and UX of the website, as it passes those responsibilities to the dealers, which I believe are more capable of handling increased server traffic.
But there are caveats to this too, one is that it would need for Golden to inventory and sort the seconds stock that will be shipped to each dealer. I don't know how much additional time and effort this will add to the company.
Another thing is that the online seconds sale was modeled after the physical onsite seconds sale, which happens specifically only one site, at Golden. Going by this, Spyderco might want to retain holding seconds sales thru their website as an online mirror to their actual sale at the main factory.
