詳細検索

AWS re:invent2016 Key Note --- Day1 ---

Avatar
by 赤川蘭丸
3 min read

AWS re:invent2016 Key Note --- Day1 ---
Translated from 日本語 • View original

Hello.

This is Akakawa, who is in charge of infrastructure.

11/30 was Key Note Day 1.

I would like to briefly write about the new information announced today.

Computing Services

Light Sail

ls VPS service for those who have never used a server. Launching an instance as easily as possible. With a fixed monthly fee including data communication amount, you can choose the service (WordPress) you want to use.

T2

A new type of T2 instance has been added, which is an instance that can emit instantaneous CPU bursts.

R4

As the successor family to the R3, the R4 was announced. The maximum installed memory is 488GB.

I3

As the successor family to the I2, the I3 was announced.

C5

As the successor family to the C4, the C5 was announced. This type is ideal for compute-intensive workloads.

Elastic GPUs

EGPU You will be able to attach GPUs to existing instances in the same way as EBS. It is suitable for those who want to run a wide range of graphics workloads, such as 3D modeling.

F1

F1 has been announced, a family of instances that can use FPGAs, programmable hardware. You can program custom hardware acceleration yourself, create an FPGA image, and run custom logic on the hardware.

Updates on DB

Amazon Athena

athena

If you find it a bit too expensive to use EMR or ElasticSearch, you can throw a simple query and analyze it. It supports ANSI-SQL, so no special skills are required.

In the future, I recommended that you build a data lake centered on S3.

Aurora PostgreSQL Ready

aurora_pg

Aurora, which was announced in 2014, will also be released with PostgreSQL-compatible version. Good news for customers who wanted to use Aurora with PostgreSQL.

AI/Deep Learning

Amazon has been using AI in the past, such as Amazon's product recommendation function. Deep learning for voice services such as Alexa & Echo.

Amazon Rekognition

reko It is an image recognition service using deep learning. Analyze the image stored in S3 and output tag information related to situation, object, and face information.

Amazon Polly

polly

A service that analyzes words omitted in text (PPAP) and generates omitted words as full (PenPineappleApplePen).

Amazon Lex

lex

What is Alexa in the first place? As a speech recognition engine, the most common way to use it is "From Prime Music? Please play a song." Developed by a third party using SKILLS. For example, Uber (taxi dispatch), Capital One (also services such as checking balance)

You'll be able to use the technology you use as Alexa.

Storage Updates

SnowBall Edge

snow

Make data <-> migration between AWS on-premises more powerful. SnowBall Edge supports capacities up to 100TB on a single HDD. Clusterable, up to petabytes. Support for NFS and S3 APIs. Lambda functions can be specified during loading.

AWS Snowmobile

track

It seems that the truck bed is a hard disk... Data can be transferred up to exabytes. By the way, it is completely waterproof! ! It seems to be only in the US so far. . .

IoT Updates

AWS Greengrass

Gg

You will be able to adopt the same AWS Plugging model on your connected devices and AWS.

It can run on a gateway, run Lambda functions, integrate with AWS services, authenticate, and integrate with peripheral IoT devices.

IoT devices use the Device SDK to collect sensor data and send MQTT messages, and status management is performed in AWS IoT's device shadows.


Managed services were announced one after another, and the foundation side was guaranteed by AWS, so I felt that it contained the message that users wanted to focus on their main business and increase business speed.

I feel that the speed of business in the United States is very fast, but I felt that in the near future, Japan will not be able to cope with the MSP services we provide, and there will be a demand for services that can help users more (which may already be the case). </->

Related Articles