Episode Transcript
Transcripts are displayed as originally observed. Some content, including advertisements may have changed.
Use Ctrl + F to search
0:00
This is episode 672 of the AWS podcast released on June 17th, 2024.
0:09
Welcome everyone to the AWS podcast.
0:11
This is Jillian Ford taking
0:13
the place of Simon to do
0:15
this intro, which I still, it never
0:18
gets old, how weird it feels, but
0:20
Simon's around. He's just on
0:22
vacation today, but I
0:24
am here with the one and only amazing
0:27
Trudy. How's it going, Trudy? It's
0:30
going really well. And thank you for that
0:32
call out. One and only amazing. I don't
0:34
think anyone's ever called me that. So that's
0:36
awesome. Oh, first time for everything on the
0:38
update show. That's right. I
0:41
probably am the only host who's
0:43
named Trudy in our whatever history of hosts.
0:45
So that might be true. And
0:47
I'm excited for the updates we have here.
0:50
I feel like you're right that it never
0:52
gets old, that it's somehow always seems hard
0:54
when Simon's not around. But I also think
0:56
we are finding our groove as when we
0:58
do this, just you and me. I think
1:01
so too. Yeah. Multi-availability
1:03
zones. We are distributed this team.
1:05
So multi-region, that's right. This one's going to be
1:07
a shorter one, but there's a lot of just
1:09
like really useful updates, I would
1:11
say for the folks who are listening. So
1:13
let's get into them. Let's start with analytics.
1:16
Amazon open search ingestion now
1:18
allows you to ingest streaming
1:20
data from Amazon managed streaming
1:22
for Apache Kafka serverless, enabling
1:25
you to seamlessly index the data
1:27
from Amazon MSK serverless clusters and
1:30
Amazon open search service managed clusters
1:32
or serverless collections without the need
1:34
for any third party data connectors.
1:38
With this integration, you can now use
1:40
Amazon open search ingestion to
1:42
perform near real time aggregations,
1:45
sampling and anomaly detection on
1:47
data ingested from Amazon MSK
1:49
serverless, helping you to build
1:51
efficient data pipelines to power
1:54
your complex observability and
1:56
analytics use cases. Another
2:00
great updates from the Open
2:02
such seen his fractional point
2:04
five open search capacity units.
2:06
The open search service can
2:09
be deployed starting actress to
2:11
O sea use. For.
2:13
Production workloads. This includes One O
2:15
Sea You which is hop asserts
2:17
capacity units for primary and stand
2:19
by indexing notes at point Five
2:22
O Sea of each and one
2:24
O Sea a total for search
2:26
across two Point Five O Sea
2:28
You active replica notes in separate
2:30
of zones. Open. Source
2:32
service will automatically skill. Up the O
2:35
C use be some workload demand. Additionally,
2:37
for death and test workloads that
2:39
don't require highveld only open search
2:41
service. Offers a one o sea
2:43
you deployment option further cutting costs
2:45
and half with Point Five Oh
2:47
Sees for indexing and Point Five
2:49
O Sea You for search. Last
2:53
one on analytics is. We
2:55
are excited to announce any streaming
2:57
job mode on Amazon Am are
2:59
surplus. Enabling you to continuously
3:01
analyze and process streaming data. Are
3:05
right next up. We have
3:07
a few updates under application
3:09
integration. Eight of Health Imaging
3:11
now supports his injuries and
3:14
architectures by sending event notifications
3:16
to Amazon Event Bridge. Amazon
3:19
time stream for live analytics
3:21
is now an Amazon Even
3:24
Bridge. Five Stargate. Simplifying
3:26
the ingestion of time series
3:28
data from sources such as
3:30
Amazon can assess Amazon Dynamo,
3:32
Db, Amazon Ask Us and
3:34
more. Pipes provide a
3:36
fully managed experience, enabling you
3:38
to ingest time series data
3:41
into time stream for live
3:43
analytics without the need to
3:45
write undifferentiated integration. Court always
3:47
loved that when eight of
3:50
as does the heavy lifting
3:52
of writing undifferentiated good. totally.
3:54
Eight of us are fabric. And. Know
3:57
court service that quickly degrades
3:59
with. That applications to
4:01
enhance an organization Security
4:03
for shirt now supports
4:05
jump cloud. Up
4:07
Fabric provides aggregated and normalized
4:10
ordered logs from popular Sas
4:12
applications such as Slack, the
4:14
Zoom Salesforce at machines you
4:16
to suit, so on and
4:18
so forth. By. Centralizing
4:20
Sas application data. Up
4:23
Fabric Health Steams gain greater
4:25
visibility into one that abilities
4:27
in a customer Sas environment,
4:30
enabling them to monitor. Threats.
4:32
More effectively and respond to
4:35
incidents. Id.
4:37
And security team and no longer
4:39
need to manage find to find
4:41
Sas integrations that take time away
4:43
from higher value tasks like standardizing
4:45
alerts are setting common security policy.
4:49
Know. We got artificial intelligence.
4:52
Amazon sees me current now integrates
4:54
model cards into model registry, making
4:56
it easier for customers to manage
4:58
governance information for specific model versions
5:01
directly a model registry and just
5:03
a few clicks. Amazon
5:07
Cute developer launches a I powered
5:09
inline completions in the command line.
5:11
As developers type in their command
5:13
line, you developer will provide real
5:15
time a. I generated code suggestions
5:18
for instance, if a developer types.
5:20
Back. Ticket Get back Tix
5:23
Q Developer might suggest back
5:25
Tic Push Origin Main Basic.
5:28
That's. A symbol back to not actually
5:30
reading tactic, just. For. You
5:32
all else developers can accept the
5:35
suggestion by simply pressing the the
5:37
right arrow. Amazon.
5:39
Titan text embedding version to with
5:42
a. New embedding model in the.
5:44
Amazon Time family of models is
5:46
now available for use with knowledge
5:48
pieces for Amazon Bedrock. Using Titan
5:51
text to embedding speed to customers
5:53
can embed their data into effect
5:55
or database. And use it to
5:57
retrieve relevant information for. Tasks such
6:00
as questions and answers,
6:02
classification, or personalize recommendations.
6:05
Amazon Text embedding to be to
6:08
his office for retrieval augmented generation.
6:10
And is an efficient model ideal
6:13
for high accuracy. Retrieval tasks
6:15
at different dimensions The mall
6:17
support flexible. Embedding sizes and
6:19
maintains accuracy at. Smaller dimension
6:22
sizes. Hobby to reduce storage
6:24
costs without compromising on accuracy.
6:27
One reducing from one thousand twenty
6:29
four to five hundred and twelve
6:31
dimensions time text embedding to be
6:33
to were teens approximately ninety nine
6:35
percent retrieval accuracy, and reducing from
6:37
one thousand and twenty four to
6:39
two hundred fifty six dimensions the
6:41
mall or of maintains ninety seven
6:43
percent accuracy. Additionally, plain
6:45
text embedding spears and to
6:48
includes multilingual support. For over a
6:50
hundred. Languages in Pre Training as
6:52
well as unit. Factor Normalization
6:54
for improving accuracy of
6:57
measuring sector similarity. Next.
6:59
Self. We have a few updates under
7:02
Compute. We. Are excited
7:04
to announce the availability of
7:06
Amazon! Easy to instance type
7:08
find are enabling you to
7:10
select the ideal easy to
7:12
instant types for your workload.
7:14
This uses machine learning to
7:16
help customers make quick and
7:19
cost effective selections of instance
7:21
type before that provision that
7:23
workloads. So using the aid
7:25
of this management console, customers
7:27
can specify the workload requirement
7:29
and get really good recommendations
7:31
on what instances. To use. Incensed.
7:34
I find or is integrated with
7:37
Amazon Q which will allow customers
7:39
to use national language to specify
7:41
requirements and then get the instant
7:43
family situation. So. This
7:46
one's really exciting because Easy
7:48
To has more than seven
7:50
fifty instance types and this
7:52
kind of functionality with the
7:54
instance type finder will enable
7:56
customers to easily choose the
7:58
best option for. They were to
8:01
requirements whether they're trying to optimize
8:03
performance or lower costs whatever that
8:05
might be and by using the
8:07
easy to and since I find
8:09
or in Amazon, Q and other
8:11
console experiences, customers will be able
8:13
to make really informed decisions about
8:15
what the best instant types are
8:17
for their workloads and basically speed
8:19
up at eight of the development
8:21
instead of having to lose all
8:23
his research themselves. find or will
8:25
help guide your decisions. I'm really
8:27
excited about this one's easy to
8:29
insensate. Find. Or is available in
8:31
all commercial eight of Us regions.
8:34
No one thinks you're That is
8:36
interesting is that we already have
8:38
a similar functionality that you know
8:40
customers may be familiar with which
8:42
is called the Computer Optimizers. So
8:45
what do you think about that
8:47
When would customers use one horses
8:49
the other? And what has your
8:51
experience been with with Compute optimizers?
8:54
Such. A third call out on
8:56
compute optimizers so of folks were listening
8:59
who don't know a computer optimizers is
9:01
it will run analysis on your easy
9:03
to instance utilization in your a that
9:05
the U S environment. so this is
9:08
great when you are running easy to
9:10
instances and he wanna be able to
9:12
understand hey. How. For utilization
9:14
said I B M I using the
9:16
right instance type so they may be
9:19
can I use a smaller one or
9:21
their savings plans opportunities So that is
9:23
great for easy to instance workloads where
9:25
you're already running them. So when I
9:27
love about this is now there's really
9:29
no excuse because it has got a
9:31
new workload in your like on I
9:34
What's going to pick a winner? What
9:36
size depicts You've got the newest his
9:38
recommendations and then if you've got one
9:40
already you've got. To. Pew optimizers for
9:42
those existing applications that are using Easy
9:44
to instances. This something for pretty much
9:47
anyone who's using Easy to as load
9:49
everything Easy Do right? Yeah,
9:51
absolutely something for everyone and easy
9:54
to sort of provides the broadest
9:56
and deepest set off instances and
9:58
these to function. These help
10:00
customers navigate that really broad
10:03
and the portfolio. To. Help
10:05
find the right one for them. All
10:08
right. Next up, we have
10:10
it up: Louis: Elastic Beanstalk
10:12
or which now supports.net Aid
10:14
on a L. Twenty Twenty
10:16
Three Elastic Beanstalk Environments. And
10:19
eight of us batch now
10:21
offers the job queue snapshot
10:23
feature enabling customers to observe
10:25
the jobs at the front
10:27
of that use. This feature
10:29
provides visibility to the existing
10:32
eight of the this batch
10:34
fair share scheduling capabilities. The.
10:36
Job Queue snapshot display of the jobs
10:38
at the front of your job to
10:40
use to assist. Administrators. Now.
10:44
On Cassandra Edmunds, Amazon Can
10:46
Act now provides time zone
10:48
support for forecasts. Amazon Connect
10:51
Agent Workspace now features. And
10:53
up the the user interface to
10:55
improve productivity and focus for your
10:57
agents. Now.
10:59
You have databases. Amazon.
11:02
Or on my sequel, three.zero
11:04
Seven. Is generally available. Eight.
11:06
It is Dms now supports Babel
11:08
Fish for over a post grass
11:10
as a source. One.
11:13
Quick update on developer tools Amazon
11:15
Cool Catalyst now supports get hub
11:17
cloud. Source Code with Blueprints Lupus
11:19
are used by Code Catalyst to
11:21
set up a code repository with
11:24
a working simple app to find
11:26
cloud infrastructure and run pre configured
11:28
see I see the workflows for
11:30
your project. Now customers can
11:32
use code Carlos. Blueprints to create
11:34
a project with a source repository
11:37
and simple source code, Si Se
11:39
workloads and build and test reports
11:42
and integrate issue tracking tools. Know
11:45
you've got to updates on front
11:47
end web and mobile. Amazon.
11:50
Apia gateway now enables customers.
11:52
to increase their integration time out beyond
11:54
the prior limit of twenty nine seconds
11:56
i know this is really exciting if
11:58
you've got any those Long-running workloads,
12:00
maybe there's like a generative AI workload
12:02
or there's some other SaaS type of
12:04
application where you're relying on a third
12:06
party and it requires this
12:09
longer integration timeout. This is definitely one
12:11
that you want to make sure that
12:13
you bookmark as update. That's going to
12:15
definitely help you. And
12:17
the last one on the front
12:19
end web and mobile updates
12:21
is Amazon Location Service launches
12:23
enhanced location integrity features, which
12:25
offer tools to help developers
12:27
evaluate the accuracy and authenticity
12:30
of user-reported locations. With
12:32
enhanced location integrity features, customers can
12:35
now use predictive tools that
12:37
anticipate user movements into or
12:39
out of customer-specified areas using
12:41
criteria like time to breach and
12:44
proximity to enhance monitoring and security
12:46
measures. Now
12:48
one quick update on the
12:50
Internet of Things, AWS IoT
12:52
Device Management as a unified
12:55
connectivity metrics monitoring dashboard. Next
13:00
up, we have Management and Governance.
13:03
Amazon CloudWatch now supports
13:05
AWS CloudTrail Data Event
13:07
Logging for the GetMetric
13:09
data and the GetMetric
13:11
widget image APIs. Amazon
13:14
CloudWatch Logs Livetail now
13:16
supports streaming CLI making
13:19
it possible to view, search, and
13:22
filter relevant log events in real
13:24
time. You can now view
13:26
your logs interactively in real time as
13:29
they're ingested via the
13:31
AWS CLI or programmatically
13:33
within your own custom dashboards
13:35
inside or outside of AWS.
13:39
AWS launches AWS Tax
13:41
Settings API, a new
13:43
public API service that
13:46
enables customers to programmatically
13:48
view, set, and modify
13:50
tax registration information and
13:53
associated business legal name and
13:55
address. This launch
13:57
allows you to automate tax registration.
14:00
updates as an enhanced
14:02
offering to the AWS tax
14:04
settings page. Up
14:06
until now, customers managing tax
14:08
registration information could only update
14:11
this information from the tax
14:13
settings page on the AWS
14:15
billing console. This
14:17
API now enables customers to
14:20
automate setting their tax information
14:22
while creating bulk accounts instead
14:25
of manually setting tax registration
14:27
information for accounts. This
14:30
programmatic supports allow customers to
14:32
build automation around setting and
14:35
modifying tax registration information. Customer
14:38
creating accounts using the AWS
14:40
account creation API and other
14:42
AWS services can now fully
14:45
automate their account creation process
14:47
by integrating the tax registration
14:49
workflow into their overall programmatic
14:52
account creation process. One
14:55
update on migration and transfer. AWS
14:58
Transfer Family increases message size
15:00
and throughput limits for the
15:02
applicability statement protocol. And
15:06
last of the update show, we've
15:08
got two updates on security, identity,
15:10
and compliance. Amazon Inspector
15:12
container image scanning is now
15:15
available for Amazon Code Catalyst
15:17
and GitHub Actions. And
15:19
we've got an update from Audit Manager, which I
15:22
think is gonna be tied with the
15:24
EC2 instance new recommendation update,
15:26
I've gotta say. Audit Manager, I
15:28
love this one because there's so
15:30
many industries, so many different types
15:33
of businesses that can really benefit
15:35
from Audit Manager. So this one's
15:37
a really good one. So AWS
15:39
Audit Manager hasn't introduced a common
15:41
control library that simplifies the process
15:43
of automating risk and compliance assessments
15:46
against enterprise controls. This new
15:48
library enables governance, risk, and
15:50
compliance teams to efficiently map
15:53
their controls into Audit Manager
15:55
for evidence collection. The
15:58
new common control library... provides
16:00
predefined and pre-mapped AWS
16:03
data sources, eliminating the need
16:05
to identify which AWS resources
16:07
to assess for various controls.
16:10
It defines AWS Managed Common
16:12
Controls based on extensive mapping
16:15
and reviews by AWS certified
16:17
auditors, determining the appropriate data
16:20
sources for evidence collection. With
16:23
this launch, Audit Manager will also
16:25
deliver more evidence mappings for
16:27
controls, including 140 newly supported API
16:29
calls for
16:32
additional evidence. You can
16:35
customize and update all evidence mappings
16:37
as appropriate for your objectives. The
16:40
library also reduces the need
16:42
to implement different compliance standard
16:44
requirements individually and review data
16:46
multiple times across different compliance
16:49
regimes. It identifies
16:51
common requirements controls,
16:53
helping customers understand their
16:55
audit readiness across multiple
16:57
frameworks simultaneously. As
17:00
AWS Audit Manager updates or adds
17:02
data sources, such as
17:04
CloudTrail events or API calls
17:06
or any newly AWS Config
17:08
rules or maps additional
17:11
compliance frameworks to the common
17:13
controls, customers automatically
17:16
inherit these improvements. This
17:19
removes the need for content updating
17:21
and provides the benefit of additional
17:23
compliance frameworks added to the Audit
17:26
Manager library. Shruti,
17:28
that might be a world record of
17:30
the fastest AWS update show we have
17:33
ever done. I
17:35
think so. I think this was a quick and
17:37
short one, but it did have some interesting ones,
17:39
as you said. Yeah, totally.
17:41
EC2 recommender for new instance
17:43
types, definitely this Audit Manager.
17:45
I really liked that one.
17:48
Shruti, where can people get a hold
17:50
of you on the internet? Yeah,
17:53
you can find me on LinkedIn, Shruti
17:56
Koperkar, or you can also find me
17:58
on Twitter. Same thing, Shruti Koperkar. And
18:01
you can also reach out
18:03
to us on aws.amazon.com/ podcast.
18:06
We've got a big button on the
18:08
page where we would love for
18:10
your feedback. And if you fill out that
18:12
form, we're actually going to read it. Like
18:14
Shruti and I are going to sit there
18:16
and actually read it because we love hearing
18:19
feedback from customers. So please reach
18:21
out if you want to hear
18:23
any new topics, anything feedback related
18:25
about the update show, the AWS podcast.
18:27
Or a topic that you want to
18:29
hear. Or a topic, yes. Please
18:32
send suggestions about what you'd like
18:34
to hear more about. As Jillian
18:36
said, we are an extremely customer
18:38
obsessed company. And everything we do
18:40
here is working backwards from
18:42
customer needs. So if there's something you want
18:44
us to do a deep dive on, let
18:47
us know. That's right. All right,
18:49
everyone. Until next time, keep
18:51
on building.
Podchaser is the ultimate destination for podcast data, search, and discovery. Learn More