91–100 of 120 posts
Earlier quoted context omitted.
Time to create a competitor, right?
There is quite a few. One I use and like is: https://sucuri.net that does security + CDN. But Incapsula, Akamai, Fastly as are good competitors too.
Yeah, at this point, we honestly and genuinely need large CloudFlare competitors. So hopefully one of their competitors is able to get more competitive.
Is Google's Cloud DNS product a competitor? https://cloud.google.com/dns/ Doesn't seem like it has the "firewall" capabilities of CF. Also found this: http://alternativeto.net/software/cloudflare/ Because of the size of the WordPress market, Sucuri.net might be a legit competitor.
Akamai and Incapsula are the only real competitors and, tbh, none of them are particularly "good" which is why CloudFlare has been so successful.
Sucuri is fine for a corporate blog or the like but it just really isn't usable for a high traffic site with alot of dynamic content.
If your website is well-coded and administered, does CloudFlare offer any performance benefit? (leaving aside security for now) If a page is static, then CloudFlare can cache it. But if you set your cache headers appropriately, and use efficient serving code like nginx, I imagine serving static content is pretty darn cheap. If a page is dynamic, then how can CloudFlare really speed it up? You don't want them serving…
As for dynamic page caching, CloudFlare offers a service called Railgun that only sends the diffs of a page when it's been changed, rather than the full page, and then re-hydrates it at the edge of their network before handing it off to end-users. Theoretically this would reduce network time by sending less traffic inside the network. I've never personally used it so I can't vouch for it, but it sounds neat.
Earlier quoted context omitted.
Wouldn't any large enterprise client like this already have a CDN or DDOS-proof hosting provider?
I don't know for sure, but they likely have to provide a private key to these providers. Other big disadvantage of a "DDOS-proof hosting provider" is that they likely are not distributed, so absorbing large DDOS is hard.
What you want is DDoS resistance which is as strong as possible, and sufficiently strong for your application.
It's essentially impossible to have a single location anywhere in the world which is able to resist the largest attacks today (many hundreds of gigabits per second); even if your site somehow has 10Tbps of interconnects, the networks near you will get saturated. (And I don't think you could buy 10Tbps of interconnect to various networks at a single building today)
You could accomplish bigger-than-single-site DDoS resistance in three ways:
1) Direct private connections to users (there are systems which are unsuited for the public Internet!)
2) Buying lots of long-haul off-Internet transport to networks around the world; running your own private network which interconnects in-core in various important networks around the world.
3) Building a highly redundant global distributed front end which then backhauls traffic (over public networks, private networks, some combination), and potentially handles some or all of that traffic at the edge directly (cache, and maybe someday edge processing...)
(#3 can also be accomplished with an application having a loosely coupled front end; if you're just serving static content this is trivial. If you can deal with eventual consistency, it's also relatively easy. It becomes incredibly difficult for regular web apps or anything with lots of writes which need to be globally atomic. There's a lot of great CS work to be done, and then tool development, to make truly distributed web apps possible, then easy...)
#1 and #2 are incredibly expensive as the number of users scales up. #3 is what ~everyone does, even the biggest DDoS targets. CloudFlare and other providers are good at doing #3 for you so you don't have to, but theoretically a large company with a lot of money could do this themselves. It's probably not a core competency of a large financial institution to do this, so modulo trust issues, it makes sense for them to pay a provider to do it. We're working particularly hard at technical solutions to the trust issues, cost, etc. to make it a good solution for anyone.
Earlier quoted context omitted.
In tech it is well known to be the opposite of that. Low end products move upward. It's very hard for high end focused companies to move downward.
Counterpoint: uber, but yes I hear you.
This is more about big established companies. It's not really even technical; it's protecting their revenue, employee/management structures, etc. built around high-end from low end.
It would be interesting to find organizations which did historically build super high end stuff which moved down into consumer. Both successful and unsuccessful examples.
iRobot MIGHT be an example -- moved down from DARPA and Packbot to the Roomba; but they weren't really established, and those were different sectors.
Apple has sort of moved downmarket from computers (and, expensive ones) to cheaper ones to phones. But those were mostly adjacent sectors, not in their core product.
I'd love some counter examples where it was within the same product, an established company (say, 10+ years selling the old product), and moved in at the low end.
How does CloudFlare compare to Akamai? Are they direct competitors?
Yes, direct competitors on the Enterprise space. On the low/mid end, only https://Imperva.com and http://Sucuri.net compete with CloudFlare.
Silly question: How does CloudFlare make any money? Any random can put any site behind their (very fully featured) free services and get free CDN, free antimalware, and free $other_services, with no seeming limits as to the amount of traffic you get. This has no impact on the target site. There are no ads or any other such. Their enterprise products only offer a few more features at a massive cost hike. How are the b…
They're an MITM service. They see your encrypted traffic in the clear. There has to be some way to monetize that. Why else would Google buy in?
Earlier quoted context omitted.
Counterpoint: uber, but yes I hear you.
Uber's moving down from Uber Black to UberX, Pool, etc., but wasn't really an "incumbent" (they are young), and it was fundamentally cheaper-than-limo even from the beginning (by letting you pay by the ride vs. per time block). This is more about big established companies. It's not really even technical; it's protecting their revenue, employee/management structures, etc. built around high-end from low end. It would b…
Same deal with SpaceX.