I'm doing a research project on middlebox appliances such as proxies,
WAN optimizers, and firewalls. Middlebox appliances are any
networking-related hardware other than routers and switches. I'd like
to learn a little bit about how middleboxes are used in real world
deployments in enterprises. Vendors often engage in surveys of this
type - but the research community knows less than we'd like to about
typical concerns in an enterprise network.
Some promises:
(1) If you give me your email address, I will not give it to anyone
else, nor will I add you to any annoying mailing lists.
(2) If you mention the name of your organization, I will not share it
with anyone else.
(3) If I publish any data from this, statistics will be reported in aggregate.
(4) I will not share the raw data from this survey with anyone other
than my advisor, Professor Sylvia Ratnasamy
(sylvia@eecs.berkeley.edu).
Feel free to skip questions and please provide approximate answers if
you have them.
Finally, to thank you for your time, we'll enter you in to a lottery
for a $100 Amazon gift card; we'll select two people to win and
contact them on November 16. Thank you!
If you have any questions or concerns, please contact me.
I find it interesting, to say the least, that all of the communication
that you have about a Berkeley research program while your email came
from washington.edu?
Thanks,
'Ayo
..... Success is getting what you want, happiness is wanting what you
get - Ingrid Bergman
... the sky is too low to be my limit ....Sent from my iPhone
I should've guessed that you guys, of all people, would notice the
discrepancy.
I used to be at the UW; I registered for this list using my UW email
address. Rather than re-register in order to be able to post to the
list, I just sent from my old email address.
I should've guessed that you guys, of all people, would notice the
discrepancy.
I used to be at the UW; I registered for this list using my UW email
address. Rather than re-register in order to be able to post to the
list, I just sent from my old email address.
I've met Justine and can vouch for her serious, laser-focused interest in middlebox research, and that if she's not really attending Cal then she's bamboozled a whole heap of CS profs over there.
But seriously, if you can help her ascertain real middlebox use cases she wants to help improve that segment of networking via useful research, nothing more or less.
Would love to see the results, although it definitely is catered more to enterprise than ISP (where many of these are probably used more than in enterprise).
It's missing a small datapoint. What types of failures are most likely to occur?(Physical/electrical, Misconfiguration, Overload)
But seriously, if you can help her ascertain real middlebox use cases
she wants to help improve that segment of networking via useful research,
nothing more or less.
Would love to see the results, although it definitely is catered more to
enterprise than ISP (where many of these are probably used more than in
enterprise).
Unfotunately ISPs are deploying many middle boxen, frequently in series,
for various reasons...cough cough cgn.
Given that these middle box infested ISPs are supposed to be providing
"internet", that seems like more fertile research grounds, as the
definition of internet is starting to shift ...at least imho.
Cb
It's missing a small datapoint. What types of failures are most likely to
heh, Until IPv6 is a mainstream, I don't think wireless companies (and soon wireline) have much choice on CGN. I believe there are plenty of CGN products that handle as much or more pps than my Juniper MX960 does. My last DDOS killed the egress pps on 2 of my NSP transits. Neither could send 2Mpps of traffic to me (ie, neither was line rate at 43bytes).
I'm confused as to the 6to4 gateway state. Last I checked, all my 6to4 is stateless.
My load balancers are also stateless.
IPS can be deployed sidelined with hardware packet mirroring and remote updates to router ACLs.
I recognize that ISPs may not keep DDOS in mind and reduce state when possible, but there is current tech that can limit state and still deploy the same services. CGN is the exception to the rule, and I've yet to see a way around it in a depleted IPv4 Internet (but as stated, most CGN is designed to handle state to the same performance levels as current router tech).
I'm confused as to the 6to4 gateway state. Last I checked, all my 6to4 is stateless.
Depends upon the technology being used. I probably should've used a different term.
My load balancers are also stateless.
Most are not, or aren't configured to be so (i.e., most seem to be set up to handle the outbound server-to-client comms, too). I see them go down like tenpins in trivial DDoS attacks all the time.
I should've guessed that you guys, of all people, would notice the