Thank you for having me today, My name is Krishnan Mani, I am a solutions Architect
with Amazon Web Services and I work closely with Mithi Software solutions, who are one
of our trusted partners who provide a service to their end customers and leverage the services
and infrastructure that Amazon Web Services provides to our many customers here in India
as well as globally.
So today I want to spend a little bit of time, maybe the next 10-12 minutes talking about
a few of the aspects.
You know just about, for example in terms of storage, the most critical aspect, in terms
of email backup and archival is that Vaultastic makes a promise to their customers such as
Mahanagar Bank that once a copy of an email is journaled and/or archived, with Vaultastic
they can pretty much rest assured that this email, the data in that email will not be
tampered with, it will never be lost and that it can be retrieved or searched for pretty
much at any time in the future.
SO for them to be able to make this promise, to their customers they need to make sure
the building blocks they choose for the applications and the service that they have created can
give them the similar guarantees in terms of the durability, the availability and the
scalabitlity of the underlying services.
On this slide if you see, typically a customer, end user has many uses for their data.
IN the case of Vaultastic, we are talking specifically around backup, restore and recover
as well as archive and of course with ediscovery you are also trying to maybe, you want to
be able to retrieve critical email, you want to meet any regulate compliance requests and
so on.
So, if you look at the storage platform and solution that AWS provides like, I mentioned
there are some building blocks that our customers can use when they create products and solutions
like Vaultastic.
So as far as storage is concerned, we have these services including Amazon S3 and Amazon
Glacier, these are what are called object storage services.
We also have other services such as Amazon EFS which is a shared file system service.
We have block storage including Amazon EBS and EC2 instant store and then a number of
other services such as Snowball, which Ashok spoke briefly about, where, when you want
to transfer large volume of data, for example anywhere between 50-80TB of data, it is no
longer sensible to rely on doing this over the internet.
So the Snowball is a service which actually you can use in that scenario.
With this background, I want to also talk a little bit about contrasting how the use
of a service is fundamentally different from the use of a particular storage setup or storage
that you may have created either on premise or in a data centre hosting that you may have
chosen to run your applications in.
So I want to talk a little bit about durability and S3 is one of the key services that the
Vaultastic product uses as Ashok said, the Vaultastic product currently is receiving
and saving close to 5Gb of data per second, which is a phenomenal data rate, and how do
you achieve the durability so that the figure on the screen, I know that it is difficult
to read this, but we refer to it as 11 9's of durability.
So when we say 11 9's of durability, I understand that it can be difficult to comprehend, but
so that you can make some sense of what this number means.
What it means that if you storage let's say 10,000 objects, so an object in S3, you
can think of it roughly as a file, or a part of a file.
So a file can become an object in S3.
It can be a small object or a larger object.
So when we say 11 9's of durability, what is means statistically, is that if you were
to store 10,000 objects or files in S3 then you can expect to lose maybe one object of
those 10,000 in about One Crore years, that is what this number really means and that
is the kind of expectation of Durability that Amazon tried to work towards.
We are basically promising our customers.
That we are investing in all of the engineering efforts and the infrastructure provisioning
so that we can make sure that we fulfill this promise.
Now I want to talk about why it is that only some services can give you this kind of durabilty
and
this level.
What we see here is, on the right hand side what we see now, Let's say you are operating
on premise, which is where you might have begun with your operations and you are currently
using some form of storage, in this case, let's say tape storage.
It has been established that with the best practice setup you can achieve something like
4 9's of durability that means 99.99% durability.
Now what do you do?
You realise maybe this is not necessarily adequate, because you may have realised over
a period of time that you tend to lose some data, because let's say the tapes went bad
or drives went bad or you could not trace where some data was backed up, so what you
then do is, you add an additional offsite storage location, which means that every time
you backup some data, you have a scheduled mechanism whereby a copy of that data is also
stored in a separate location, maybe 50, 100, 200 miles away, so that it gives you a better
chance to recover the data in case some of it is lost at one of the locations.
Now it has been established again following best practices, this is only going to give
you another incremental duarabilty, which is 5 9's of durability, now going ahead,
when we architected a service like S3 or Glacier, we are deliberately investing in the engineering
behind this service, basically you are jumping from what you can achieve on premise, or what
you can achieve on premise and off site copies and so on.
However it can never give you the kind of durability that a service which is deliberately
engineered to give you the quality can give you which is you are jumping directly from
a 5 9's to an 11 9's kind of durability.
How do we achieve this?
So whenever a customer suggests, when the Vaultastic product is writing data to S3,
as S3 receives objects, it is basically persisting copies of objects to multiple facilities within
our data centres, and as part of this service we are constantly verifying that the integrity
of the data n these multiple locations and as and when we upgrade our infrastructure
as and when we attend to any hardware, software failures and as and when we upgrade our own
software ans services, we take special steps to ensure that these number of good copies
of your data continue to be maintained, continuous to be accessible to you at the service level
that we are promising to customers.
So these kinds of things, this can only be achieved only if you use a service concept,
so if you think about it, Mahanagar Bank is providing banking services, Mahanagar Bank
wants to focus on giving the best banking facilities to its customers They are relying
on Vaultastic so that Vaultastic can focus on giving the best backup and archival services
and in turn Vaultastic is relying on AWS and the S3 service and we in turn are focusing
on giving you the best durability, availability and scalability.
So in this increasingly complicated and fast moving world, we believe that this is the
kind of model where we are all able to co-operate and work together and scale how we can best
serve our customers.
Here are some of the large users of this service on a global scale.
Many of these companies and names you might be familiar with because they act as multinational
corporations such as Philips.
Little bit more around S3, so at this point in time, the S3 Service has been in operation
for over a decade, we started sometime back in 2006, ever since then we have an increasing
number of customers storing billions of objects, or you can think of objects as files and the
data that we store in the service is growing rapidly, it is both durable and reliable.
We are constantly looking to ensure hat it can provide high performance at a massive
scale, so it never gets easier for us as you can see, as more and more customers with more
and more data are adopting this service across the world.
So for example, to talk about the performance, Mithi Software or Vaultastic never needs to
do anything special as the amount of data that they are saving to S3 is growing every
second.
They don't need to provision or request a particular storage amount or a rate at which
they can read and write the storage, the service is designed in such a way that it can respond
to increases in the rate at which the data is being read, the rate at which the data
is being stored and so on.
So we are doing this for Vaultastic and all our customers at a massive scale.
In terms of Data transfer, what happens is that, the Vaultastic product actually uses
our API's for S3 and translates the operations that they perform to be able to save email
or to discover email into the API operations, so this makes it easy and flexible, because
you can test this.
You can make sure that this works well, these operations can be monitored, so that both
Vaultastic and us, we can basically ensure that we are able to meet the customer service
requirements.
S3 is secure by default, so every customer has a number of different options and that
is available to them, so when you store objects in S3, the Vaultastic product and Mithi Software,
they make decisions on a granular level, right down to a single object or file.
They can determine who is allowed to read that data, who is allowed to write that data,
who is allowed to even detect the presence of a particular file or object, so they are
able to do this on a granular level across their customers, even when they are operating
in a Software as a service fashion.
SO in this way they can ensure that there is good isolation and they are able to assure
each of their customers that their data will be maintained safely, maintained independently
and that only the particular customer will have access to their corresponding data.
S3 is gradually adding many different, new features as we innovate, so for example as
of last year we have introduced a service whereby data can be queried in place.
The reason why customers and you can start to take advantage of these innovations, going
forward so that both the Vaultastic product and the end users can rely or find the new
ways to take advantage of the rapidly increasing data.
The challenge today for customers, once you have gone past the point of how do I store
the data securely, once you have conquered that challenge, the next thing that you are
thinking about is that, now how do I actually take advantage of all of the information that
is, there is a wealth of information in this data.
Also S3 is just one of the services that you can choose to use and it is integrated with
an ecosystem of other AWS services at last count, over a 100 plus services including
compute, networking, databases, analytics and so on as well as an ecosystem of partners
just as Vaultastic and Mithi Software solutions, there are many other partners as well, who
are building on top of the AWS cloud so that you can actually get more value out of this
investment.
So that was pretty much, basically I wanted to just try and talk a little bit about one
of the key services that Vaultastic uses and if you have any questions, please do communicate
that to Ashok and I will be happy to answer them later.
Thank you very much.
Không có nhận xét nào:
Đăng nhận xét