So what are some of the concepts and one of some of the methods that we can use in order to make maintain this high up time? Well, the second that look in another one of our concepts real quick, and this is going to be the concept of high availability Now, high availability is essentially having a strong up time having riddle,
having little time that our network is down called downtime.
Now the amount of our availability that we need to ensure is going to be based off of again our agreement with our management or R S L A. That determined that lets us know what are up time is required to be an high availability in up time in general is going to be measured in percentages.
Now, this percent of these percentages essentially measure the percentage of
the percentage of the time that our network is up the percentage of the time that our network is available. So let's take a look at some of these percentages
now if we tell somebody or if we have an S l. A. That says that we're going to have a 90% network up time, that means that we have an allowable 36.5 days per year that our network can be down, or that our devices are service's could be unreachable.
So 90% up time is going to be 36.5 days per year. Same with your Internet service provider if they so that your net, if they say your Internet is going to be dead, is going to be available 90% of the time. That means it's going to be available 36.5 days per year. Now again, this is based off of 24 hour day, seven day week calculations,
So don't think that 90. But this is, 0 90% of
if they say they want 90% up time on a 40 hour week that that's throwback. 36.5 days per year, 16.8 hours per week. No, this is based off of a 365 3 This is based off of a 365 day calendar, 24 hours a day, seven day a week type idea.
us. And think of this like your Internet service provider because they're 24 hours a day, seven days a week, the 65 days a year
except on leap years.
If they tell you you have a 90% availability rate, then it's going to be down on Lee 36.5 days per year. That's the mat best the most that we can allow for that 90% up time. This equates to 16.8 hours a week.
So if we tell somebody that we're going to have an up time or we're gonna have enough time of 90%
then that means that we could be down at most 16.8 hours per route.
Next, we moved to 99%. 99% is going to be 3.6 days per year and is going to have a down time of that most 5.4 hours per week
and then lastly, up here. Anyway, we have 99.9% up time, which is 8.67 hours per year
and 10 minutes and 10.1 minutes per week.
these are just some of our These were just some base percentage is this? This will be what's referred to as 19 up time. So if you hear somebody say, Oh, are up time is we have 19 or two nines or three nines won nine is going to be 90% to 9 is 99%.
Three nines is 99.9%.
And then you may hear some people boasting Oh, we haven't up time. We have we have five nines up time. Well, that means they're saying they have Ah, 99.999% up time, which equates to less than five minutes a year down.
So if someone is asking for five nines up time, you know you have a lot of redundancy and a lot of fail overs that you're gonna have to implement. So keep that. Keep that in mind when someone asks if when someone is trying to calculate your trying to calculate how much allowable down time to get 1/2 of your network,
understand these different percentages and understand what the maximum allowed downtime is for your network
Now that we've done our high availability percentages out of the way, let's talk about some of our methods that we can use to optimize our network to allow for high availability. First, well, at first that we have are cashing engines Now. Cashing engines are essentially our devices or software
that allows our that allows our network at different points.
Thio save frequently used resource is so that they're available to other available to access more quickly than typical resource is would be now. This could be anything from low bouncers to proxy servers to even our local cash.
So what do we mean by cashing? Are saving a resource? Well, for example,
our proxy servers are content filter. Chauffeurs can act as a cashing engine for certain websites. So if we have users that are trying to access a certain website over and over and over, we have users that are trying to go to a certain website repeatedly. Whether it's a,
it's a frequently access company website. It's a frequently access resource.
Then, rather than having to go all the way out to the Internet, get that resource and then give it back to the user, that proxy server could start just saving a local copy of the data that it sent it sending to the users that way when it's requested, it all that has to do is just send its copy that it has and make sure that copy stays up to date.
This prevents this helps out in a couple of ways. This helps to improve network connection speeds if we're tryingto request something on our network and were able to get a response back from a device that's a couple of hops closer than the device that we would have to go all the way out to them. That's gonna improve our speed of connecting to that. Getting that resource.
It's gonna also reduce network strain because it's going to be
There's gonna be less distance that we have to traverse. There's gonna be less devices that have to use a resource is to service. I request, if we're using are cashing engines of where if we have our devices that are cashing, these resource is and our saving them and just giving us the local cached copy rather than having to use up network resource is to go out and get
the exact same thing just from the originating service
so we can have load balancers that also had built in cashing engines and can save can save requested service's rev in having two syndicates in the connection request to one of the servers that they load balance. They can simply just give a copy of the data that they actually have. Proxy service could do that.
And we also have local cashes on individual user machines.
For example, being s. Can we can we have a local being s cash in our computers? So rather than every time our computers, rather than having our computers every single time, they need to resolve a domain name toe I p address. Rather than having to go out to our dean s servers every single time,
they can simply use their local cached copy of the D. N S
or of the D. N s reserved being named the I pee. So you need to understand how our different devices can use cashing. And it's a very strong That's very strong functionality that can help with maintaining our up time. Because the less strain that we put on our network,
the less likely we have for those devices to essentially burnout or those devices too
run out of available resource is to service our clients. So the more we make sure that we have available up time, so cashing engines are strong resource, but
we also need to balance cashing engines. How long may cash certain resource is with how frequently those resource is update we may not want if we have. If we have Stiles on a file on a shared server that are being shared out to multiple users in an office,
that we may not want our local devices
to save a copy of those files if they can't access them on the network, especially with those device. If those files are being accessed potentially by multiple people at once, because our computer may try to go out and access to the file,
another user is currently using it. So our computer says OK, I'll just use my local cash copy of this of this file,
makes changes to the local cash copy and then tries to save the local cash copy over the other over. The network shared copy and totally messes up changes.
Windows seven in particular, has a feature where it will actually save data from your network. Drive onto your local machine called offline files, so that if your computer is not connected to the network, you can still technically access files that would be on a network shared device. Now,
if this network share is only you connecting to it,
then that's great because then your computer, you can access those files. When you're not connected to the network, you can make some changes to them, and then the next time you connect to the network, it'll just update those files. But if it's trying to create an offline cache of files that multiple people are using, then that's not so great.
This file's can change over time.
You may not have the same local cash copy as the actual copy on the server, so we may need to turn that off and it may. It may require more strain on our network, but it's a required thing that we had to turn off because we can't use that cashing engine with those shared Resource is, so we have to balance
the amount of the amount of network resource is that we allow it to be cashed
with the type of resource is that were using and we're changing all the time because if they're changing all the time and there they potentially can change while a cashing engine is giving out old copies people. That's a bad deal. We don't want people to be getting old copies of data,
especially if it's dated that they're potentially using for calculations or they're using
on. They may change and try to re save over a new a copy, so we may need to turn off cashing engines in those particular instances. So something the way against on your network when you're trying to determine whether or not to use cashing in a certain instance.