File Storage

WEBVTT
1
00:00:07.340 --> 00:00:11.640
In this video, we’re going to talk
about File Storage in greater detail.
2
00:00:11.640 --> 00:00:15.420
Like direct attached storage,
file storage must be attached
3
00:00:15.420 --> 00:00:18.960
to a compute node before it can be
accessed and have data stored on it.
4
00:00:19.473 --> 00:00:25.980
However, File Storage can be less expensive, more
resilient to failure, and involve lesser disk
5
00:00:25.980 --> 00:00:29.760
management and maintenance for you as the user
to do as compared to direct attached storage.
6
00:00:30.540 --> 00:00:35.460
You can also provision much larger amounts of
File Storage and present it as a disk to a server.
7
00:00:36.060 --> 00:00:39.000
File storage is mounted from
remote storage appliances.
8
00:00:39.597 --> 00:00:44.580
That is, the physical disks are contained in
a separate, specialised piece of hardware and
9
00:00:44.580 --> 00:00:48.960
they are then connected to the compute node via
the underlying infrastructure in the data center.
10
00:00:49.800 --> 00:00:55.200
These storage appliances are not only extremely
resilient to failure, the data is also far
11
00:00:55.200 --> 00:01:00.000
more secure in them as these storage appliances
provide services such as encryption in transit.
12
00:01:00.660 --> 00:01:03.780
These appliances are all
managed by the service provider.
13
00:01:04.620 --> 00:01:08.040
File Storage is mounted to compute
nodes via an ethernet network
14
00:01:08.522 --> 00:01:11.880
– the same kind of network that you might
receive email or browse the internet over,
15
00:01:12.184 --> 00:01:15.300
although this ethernet network is
normally dedicated to the task.
16
00:01:15.960 --> 00:01:20.111
This means it can sometimes be referred
to as ‘Network Attached Storage’,
17
00:01:20.415 --> 00:01:23.700
‘Network File Storage’ or simply ‘NFS’.
18
00:01:24.300 --> 00:01:27.660
One of the issues with ethernet
networks is that their speed can vary
19
00:01:27.953 --> 00:01:30.353
– the more loaded an ethernet network is,
20
00:01:30.353 --> 00:01:33.840
the more likely it becomes that its
speed or bandwidth will be affected.
21
00:01:34.322 --> 00:01:37.980
Of course, Cloud Providers build
their storage networks to handle
22
00:01:37.980 --> 00:01:43.140
very high volumes of traffic but even so,
consistent speed cannot be guaranteed.
23
00:01:43.695 --> 00:01:46.080
Therefore, File storage tends to be used
24
00:01:46.080 --> 00:01:49.500
for workloads where consistently high
network speeds are not a requirement.
25
00:01:50.700 --> 00:01:56.220
In terms of workloads, File Storage can typically
be mounted onto more than one compute node at a
26
00:01:56.220 --> 00:02:00.780
time, where the mounted disk or volume looks
just like another drive on the compute node.
27
00:02:01.500 --> 00:02:04.980
The ability for File Storage to be
mounted to multiple compute nodes at
28
00:02:04.980 --> 00:02:09.120
a time make it an ideal solution where
some sort of common storage is required
29
00:02:09.560 --> 00:02:14.880
– for example, a departmental file share, a
‘landing zone’ for incoming files that need to
30
00:02:14.880 --> 00:02:20.099
be processed by an application, or a repository
of files that a web service might access.
31
00:02:20.550 --> 00:02:23.220
In these applications, the potential variance
32
00:02:23.220 --> 00:02:26.405
in the speed of the connecting
network is not really an issue.
33
00:02:26.405 --> 00:02:28.500
Of course, where cost is an issue,
34
00:02:28.500 --> 00:02:33.720
you can use file storage for other applications
such as databases, but the trade-off is speed.
35
00:02:34.296 --> 00:02:36.960
When you provision file storage, one consideration
36
00:02:36.960 --> 00:02:40.560
you need to take into account is
the IOPS capacity of the storage.
37
00:02:41.280 --> 00:02:45.000
IOPS stands for Input/Output Operations Per Second
38
00:02:45.000 --> 00:02:48.758
and refers to the speed at which
the disks can write and read data
39
00:02:48.758 --> 00:02:52.740
(note this is not the speed of the network
between the storage and the compute node).
40
00:02:53.280 --> 00:02:57.360
The higher the IOPS value, the faster
the speed of the underlying disk.
41
00:02:57.873 --> 00:03:00.873
A higher IOPS will also normally cost more.
42
00:03:00.873 --> 00:03:05.743
Understanding IOPS is important because if the
IOPS value is too low for your application,
43
00:03:05.785 --> 00:03:10.105
the storage can become a bottleneck and
cause your application to run slowly.
44
00:03:10.493 --> 00:03:12.900
Alternatively, if the IOPS is too high,
45
00:03:12.900 --> 00:03:16.380
you will probably be paying more
than what you need to for your storage.
46
00:03:17.220 --> 00:03:21.900
For example, a file share may be mounted
on 30 different compute nodes and an
47
00:03:21.900 --> 00:03:25.980
application writes and requests data to
and from that share 60 times per minute.
48
00:03:26.797 --> 00:03:29.437
You can average that out
to 1 operation per second.
49
00:03:30.000 --> 00:03:34.920
With this simple example, you can see that each
application has different IOPS requirements.
50
00:03:35.700 --> 00:03:39.300
In the next video, we’re going
to talk more about Block Storage,
51
00:03:39.300 --> 00:03:44.040
how it compares with File Storage, and when
you will typically use one over the other.

File Storage

Uploaded by

Copyright:

Available Formats

File Storage

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

File Storage

Uploaded by

Copyright:

Available Formats

WEBVTT

You might also like