Shulou(Shulou.com)06/01 Report--

This article will explain in detail how to analyze the NetApp StorageGRID architecture. The content of the article is of high quality, so the editor will share it for you as a reference. I hope you will have a certain understanding of the relevant knowledge after reading this article.

In fact, NetApp's StorageGRID has a long history, and its source is Bycast.

Bycast is a Vancouver company whose product, StorageGRID, was launched in 2001, focusing on the medical industry.

NetApp bought Bycast in 2010. Up to now, StorageGRID has released 11 major versions and about half a year small versions.

Because StorageGRID has a history of 20 years, its enterprise characteristics are relatively rich, especially its policy-based information life cycle management (ILM) is the biggest difference between it and other products.

The number of customers and deployment scale advertised by StorageGRID is small for a 20-year-old product. Recently, however, IDC's object store MarketScape reported that NetApp has now enhanced the investment and promotion of object storage, and there have been several cases of 100PB.

I think supporting 16 sites, 100 billion objects, policy-based content lifecycle management and so on are the highlights.

NetApp StorageGRID, like the object storage of IBM and Hitachi, is composed of different node types.

The management node uses Prometheus monitoring, and the storage node uses Cassandra to save metadata. The API gateway node (optional) implements load balancing, and the archive node (optional) archives data to the tape library. If you need to support NAS, you need external NAS Bridge (VM deployment), support SMB and NFS, support file and object interworking, but it does not belong to the node of the StorageGRID system.

StorageGRID supports three deployment modes: all-in-one machine, virtual machine or physical machine / container, which is very flexible.

The hardware of the all-in-one machine is actually E-series mid-end storage. One of the two controllers installs StorageGRID OS and the other installs SANtricity OS.

Because the new version of SANtricity supports DDP (similar to Huawei RAID 2.0), it can implement two layers of erasure codes. The advantage is that hard disk failures can be reconstructed within nodes.

In order to improve performance, SG6060 specially added a separate server with large memory to do the controller of SG.

Pure software deployment supports both virtual machine and container deployment.

Container images can be deployed through K8s, but this deployment method is not officially supported yet.

Support for uniform namespaces for up to 16 sites

Support for cross-site EC

Latency 100ms can also be deployed

Can be extended to the public cloud environment

Two-layer erasure code is the main push data protection method of NetApp.

This approach takes into account disaster recovery, reconfiguration performance and acquisition rate.

ILM policy determines the location or medium of data storage, storage time, and data protection measures. It has powerful functions and convenient management.

Policies can be adjusted at any time.

Different ILM strategies can correspond to different service levels and meet the diversified requirements of various services.

Tenant administrators can be set up independent of system administrators to meet the requirements of hierarchical management of large enterprises.

Cloud images can make full use of the analysis capabilities of public clouds; bucket event notifications can be processed automatically; and search integration uses external Elasticsearch to achieve metadata search (attribute reverse search).

Providing a variety of solutions with partners is what traditional object storage does. However, NetApp docking Hadoop also uses a general S3A client, and does not have its own client.

Together with other NetApp products, Data Fabric hybrid cloud solutions are available.

The main purpose is to increase the high-performance SG6060 all-in-one machine, and the system performance is enhanced.

The StorageGRID support file must be a NAS bridge, and the NAS bridge has no file limit because it is only a translation gateway.

The new version supports HA using the VMware Hram A feature and can be used for cross-site disaster recovery.

The metadata of the NAS bridge is stored in the SG bucket (the bucket has the function of cross-site disaster recovery) to achieve the purpose of disaster recovery.

Now EC in ILM policy rules can support small objects of 200KB, and even smaller ones can only be stored in replicas.

S3 API support is relatively perfect, of which the green part is a new function.

Let's take a look at the third party's comments.

In the IDC report, NetAPP is in the leader quadrant. But in the Gartner magic quadrant, NetApp is in the visionary quadrant.

1. Hybrid cloud is the strength of NetApp, which supports advanced features such as IAM, bi-directional data synchronization, etc.

two。 No native support files, independent external gateways are required

All the key competencies scored more than 4 points, with a balanced score, with performance tied for first place with IBM and Scality.

The score in each scenario is more balanced, there is no obvious deficiency, and the score of hybrid cloud is higher, second only to Scality.

[summary]

Generally speaking, although NetApp StorageGRID has a heavy historical burden, it is precisely because of its long history that it is still very rich in functions. The more unique features are (especially the red function):

Policy-based ILM

Number of multiple sites: 16

Barrel synchronous replication

Cross-site EC

Cloud analysis

Container support

Platform service

Hierarchical EC

Multi-tenant

Archive to tape or S3

However, NetApp StorageGRID still has some deficiencies in the relatively new modern object storage, such as in the following features:

Extended cluster

Small object merging

Dedicated HDFS high performance client

Expansion of the whole pool

However, the function of NetApp StorageGRID is really perfect, compared with other manufacturers, there are few missing major features. However, not all features are needed by Chinese users, such as platform services, and AWS may not be very useful because of its small domestic share. Moreover, the domestic use of object storage is a lot of new application innovation, users hope and object storage manufacturers'R & D personnel to do some discussion or even customization, and now NetApp domestic teams are basically merged into Lenovo Lingtuo joint venture, the joint venture company estimated that still prefer to push products like NetApp FAS, may snub object storage, because the coordination of US R & D resources is still not very convenient.

On how to analyze the NetApp StorageGRID architecture to share here, I hope that the above content can be of some help to you, can learn more knowledge. If you think the article is good, you can share it for more people to see.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.